I'm building a machine and in the last second I changed from the 2696v3 to the 2697av4. Even if its slightly weaker in terms of raw performance, the newer architecture and supposedly better power efficiency make me feel its a better choice, especially since Ill be using a Chinese board, where VRMS are usually very weak. A comparison between those two (with TB OFF / ON on the V3) would be great!
I absolutely love this content, I've made up quite a few Xeon (x79/x99) gaming rigs for people over the years and never had any complaints. I've got a Xeon 2696v3 next to me that's going in a Dell 5810 just for fun. Its mental that these things just won't die. You can pick up an x99 Xeon + mobo for around 60% the price of an 5600+mobo second hand so the price to performance is still there.
Great breakdown! A lot of your videos remind me of a few anantech articles that really broke the lga 2011 v3 cpus down in incredible detail. I remember they all remarked on how the v4 series pretty much just held the line at best and often reduced the avx clocks a little on all core loads. The ring bus core count also went up from 8 to 10 to compensate, however. Looks like you were hitting a pretty solid 3.4-3.5 ghz avx with the v3 I bit the bullet and decided to go for a e5-2689 v4 and, while the records are incredibly in complete, I estimate it's all core avx turbo to be around 3.5-3.6. Fingers crossed. =)
@@mxsugz Kind of oddly. It's faster in single threaded tasks but only barely. The 2690 seemed to do more with a lower tdp. I suspect that the memory controller on the 2690 is a little more substantial. I base this theory off of aida64 latency tests showing the 2090's higher cache was also faster and had lower latency. Main memory access was also slightly faster on the 2090. Games like cyber Punk seem to run worse and hotter on the 2689 over the 2690. However, however. I got it for one game and that game was Elden Ring and it seemed to really help exactly where Elden Ring needed it, which was with single core speed. The main thread of Elden Ring really craps the bed if it maxes out the core it's running on. The 2689 gave me my 8% extra margin I needed to hit that limit just that much less often to make the game noticeably more fluid. So I'm happy with the chip. I'm happier still with the price I paid, $60 US and I'm still happy to have the old 2090 to fall back on in case I do something like start doing more video production.
Are you watching me?! Lol. I'm literally sitting between a Xeon e5-2697a system and a R5 5600 system. My hope is to have one Proxmox server for VMs and primary storage and an Unraid server for Docker and backup storage. Love your informative videos, in particular in-depth looks at the X99 platform in modern day scenarios.
Im leaving my experience here: I have mutliple xeons in my hand and I concluded that the best xeon for gaming is the Xeon 1660 v3. I'll mention all the points of all the Xeons I have: 1: Xeon 2690 v3 - Great cpu for the price ($9). It can mantain full frequency after turbo unlock. For games like warzone struggles. 2: Xeon E5 2696 v3 - Best Efficiency / Perfomance you can have under x99. At some points it can compete with the Xeon 1660v3 (4.3ghz). However it uses like 60-90 watts less. Recommended to undervolt and disable 8-10 cores for 3.8ghz (unlock). If you are worried of watts go with this one. Due to high amount of cache memory and freq this can handle games as warzone (110-120 fps) 3: Xeon 1660 v3 - if you pick this one you don't care about your VRM/Comsuptions or any risks involved. It's inferior to other cheap xeons when used on stock. To go around 4.5ghz / 4.4 ghz you will have to go above 1.26v in vcore. Going below 4.3ghz cannot compete with xeon 2696 v3. Since this processor only has 20mb cache. However you can overclock cache to 3.5ghz and beat xeon 2696 v3. At this point your cpu/vrm/mobo would be dealing with ton of power. With my Rx 6700 xt and Xeon 1660 v3 at 4.4/4.5 ghz I'm consuming around 460 watts! (Gpu undervolted even). Which means cpu is doing this lifting.
The main reason why I disable HT is because from my testing (conduced quite a while ago) the 1% and .1% low results are much better without it. An additional reason - I don't want to show that many threads in MSI Afterburner overlay. :D
I think one of the most interesting results to come out of this video is the study involving the 16c/16t vs 8c/16t. I may be mistaken but the only real difference I see between the shutting of hyper-threading with the one test and shutting off 8 cores with the other is the cutting in half the cache resulting in a 10% increase in frame rate with the 16c/16t test relative to the 8c/16t test.
I have a 2697 v3 coming soon, my BIOS is patched with Ultimate patcher tool (x99 - 8M) and I was wondering how the turbo boost unlock works with it, and how to properly apply a voltage offset, thanks for the informative videos
Dear Miyconst I am folowing yor channel long time now and I have one qoestion for you if you can reply, which one xeon for x79 or x99 is best to pair with gtx 1080 ti? Thank you :)
The simple answer is: i3-12100F, the complicated answer is E5-2697 V3, even more complicated answer: wait a bit until my set of V4 Xeons arrive and I test them, maybe there is a better option than E5-2697 V3.
Thank you very as always very informative I have big question i hope you'll answer me I'd rather go with your favorite setup E5 2697a v3 + 32g ram + huananzshi x99 tf Or I'll go with E5 2680 v4 + 32 ram + mr9a pro
super enlightening! Thank you for all the hard work! It's safe to say if you're building a new system you are way better off going with a ryzen 5 2600 and a b450m. Heck you can probably build a ryzen 5 3600 system for the same cost as the top of the line xeon x99 system, but the ryzen system will have way more room for upgrades in the future.
Thats a shame about your camera set up but you have an awesome fluffy cat! What is his/her name? I think I've shared this before but my cat "Woolley" looks very similar to yours and happens to sleep belly upwards like shown in here in your video. Its a shame RUclips doesn't allow pics posted in the comments section.
Great video, do you think a 2697A will continue to bottleneck at 3440 x 1440 a 7800xt that much? I do have a QD4 huananzhi will it support the processor?
Hello! Very interesting video! I have a Intel Xeon E5-2697A v4 with a RX 6500 XT and the performance are good. My problem is that I have a serious issue with a server motherboard Supermicro X10SRL-F, it's not suited for gaming and it burned my videocard. Have you any experience with server motherboard? What do you suggest for replacing it considering that I have 8x16GB DDR4 2400 ECC RAM? Thanks in advance!
Hi. If your motherboard fried your GPU then it's simply a defective unit, it has absolutely nothing to do with "gaming" or "server" motherboard. 8 x 16 GB is a total overkill for gaming, thus I am not sure why you need it. My go-to Chinese X99 motherboard is Machinist X99-MR9A (Pro).
@@Miyconst thanks a lot for the fast response. First reason is because I had it for free, second reason is because sometimes I virtualize devices and I need a lot of RAM. Can you suggest me a model with 4 ddr4 ecc slot? Thanks in advance for your content that's exactly was struggling in these months and I finally found someone with the information that I needed. I was very skeptical at the beginning with this Chinese motherboards. Keep up the good work!
how do you feel about the e5-2697 v4 vs the e5-2697a v4 debate.... i currently have the e5-2697 v4 and have contemplated getting a e5-2697a because of the higher base clock which is .2 ghz more which im thinking is not going make much of a difference. id be intrested at seeing a side by side comparison of the two all the videos i find are not in english etc
Excellent video ... But I have a huge doubt for my case friend ... I have a Huananzhi QD4 Motherboard with 32 gigs of ECC quad ram .. and I have a Xeon 2697A V4 processor currently in use on this motherboard ... but when I bought the KIt it came with a Xeon 2670 V3 Processor ... I made a mistake by putting a stronger processor ... since the one I'm using has 145 TDP and the other 120 TDP ... will I damage the motherboard in the long run ?? is that when the processor pulls I notice that the VRM heats up a lot ... with Xeon 2697A V4, but it also heats up with the Xeon 2670 V3 ... after all, where do we stand ?? I would like to hear an opinion. ??
With either of these CPUs you must have good cooling and good airflow inside your chassis. If the VRM heats up with E5-2670 V3, I can assume that you have very bad cooling inside the chassis or at least the VRM is not properly cooled. So with E5-2697 V4 you are really risking to burn it down.
@@Miyconst Thank you... for your opinion .. so in this case it's better the processor Xeon 2670 V3.. ? because the TDP is more Lower? .. actually the temperature of processor is around 32º degrees celsius of course with ( Xeon 2697A V4).. what you think?
Watching the Alan Wake DX12 benchmark-- for some reason the V4 keeps constantly downclocking its cores to 300mhz while the V3 does not 🧐 that kind of throttling might raise 0.1% lows if not introduce microstutters and/or affect the whole run.
The only v4 CPU worth considering at the moment is the E5-2689 v4, which is basically an overclocked and undervolted i7-6950X. It can be hard to find, and depending on its price, you might as well just go with an i7-6900K or 6950X, since they retain their value much better and for longer due to the "i7" branding.
Hello Miyconst, On BIOS I have option for disabling cores, for E5-2697A v4 (or in general for XEON e5 cores) do you think it makes any sense? It can improve performance or reduce significantly the power consumption? Or it's just not relevant? Thanks!
Would it be good idea to pair 7900xtx or 4070 ti Super with 2697a v4, 2697 v4, 2699 v3 ? I'm concerned regarding bottleneck. My motherboard is Msi x99A plus, support up to 2699 v4, which is much more expensive than v3. Which combination would you recommend? I'd like to play 4K with 70-120fps. Thank you for your great videos
Great content regarding e5-2600v4. I have an e5-2680v4 (14c,28t) CPU and my R23 score is 12127. Strange that your score is so low in comparison. Could it be the Chinese motherboard? I am using an Asus X99 Prime A-ll with 96GB of ECC DDR4. My system has been rock solid since I cobbled it together a year ago. Most stable system I have ever put together. it's used for scientific computational modelling.
I assume you have Hyper Threading enabled with your E5-2680 V4, right? In that case it explains why your score is higher, you got 14 Cores / 28 Threads, I tested 16 Cores / 16 Threads.
Im still enjoying my E5-2697v3. On my Rampage V extreme motherboard. Recently got a RTX 3070 so need to watch and do the Re-bar mod. I need a suggestion. Currently im runing 12c/24t and the cpu goes to 3.5ghz, any baseclock oc makes the system show no picture with the RTX 3070. (I used a Titan X maxwell woth 105mhz baseclock and the cpu went to 3.7ghz with 12c) What should i do? Keep the cpu 12c/24t at 3.5ghz 14c/14t or go 10c/20t? (10c active the boostspeed goes to 3.6ghz)
I would just keep the 12c/24t configuration, seems like the sweet spot. Resizable bar makes little to no difference with Nvidia GPUs, so I wouldn't bother.
"E7-8890 v4 is LGA2011-1 which is not compatible with X79 (LGA2011) nor X99 (LGA2011-3) boards. LGA2011-1 was only used in high-end servers with up to 8 sockets"
Miconst how does a i7 7820x compare with the 2667v4, ive seen a seller in my area offering a dell t5820 with i7 7820x about 160$, and the same guy is also listing a hp z440/dell t5810 with xeon 2667v4 for about 120$, is the newer platform better in terms of gaming, because i have seen reviews of skylake x cpus running rather hot and inefficient and also do they performance rather poorer in games, I have no way to check as I only have an old laptop, and I don't want want to buy new stuff, it's out of range for me, humbly awaiting your response
That's a very tough choice to be honest. The LGA 2066 CPUs are indeed very inefficient and bad for gaming but you get faster memory and an option to upgrade to other i7/i9. If it would be me, I would probably pick T5810 as it's cheaper and invest the savings into a better SSD/GPU/monitor.
I use the E5-2697A V4 + RX 6700XT on a Huananzhi F8 mainboaard. A set of cheap HyperX DDR3200 non ECC ram 32 GB. Granted I only play Call of Duty, it actually plays it pretty ok. I have HT enabled. I have installed the Rebar unlock but it causes some instability so have it turned off in the drivers. The 2697A V4 is a bit of a rare CPU so probably a bit more expensive compared to other E5 counterparts. Call of Duty is known to scale well with a lot of cores and likes memory bandwidth so no surprise there that it does well. I get a slightly higher score in R23 though, roughly 14.000 not sure what is different there (maybe because I use HT?) This E5-2697A system is a few years old now though, for the low cost / performance segment I would not turn to the Xeon anymore. This will likely continue as a local NAS/Visualisation server for me though and there it does the job fine.
Vega 64 is not strong enough to have any meaningful use more than 8 GB vRAM. There are already quite a few videos on this subject and the quad channel memory configuration won't change the overall picture.
@@Miyconst My attempts to see HBCC on a quad channel setup failed so far, but I take your word on that. I have such a setup at home and the memory bandwidth is huge. Most of the HBCC videos I see are limited by wimpy chipsets and narrow memory bandwidth, so I was looking for something in this specific combination. Many thanks again for the answer and Slava Ukraiini!
@@Miyconst sorry for the necro post. I've tested it and got a nice bump on Quad Channel DDR3 for E5 2680 V2. In 1440p with RX5700 I could not get more than 28 FPS in Talos Principle 2 using FSR3, with mildly overclocked RX Vega 64 I got 40+. The HBCC did its job and met some speedy memory there. VRAM usage jumped from 8GB to 10+ and the extra RAM bandwidth was observed at dynamic texture loading. On dual channel DDR4 3000 it was sensibly slower than Quad Channel DDR3 1866. Not to mention that for the price of 16 GB desktop DDR4, one can get 64 GB server DDR3.
@@cripplefpv7354I can't add more, whether Xeon E5-2697 V3 is better depends on your i7's overclocking level and what games you play, also whether you apply or don't apply TBU.
Comparing a turbo unlocked v3 to stock v4 is kinda apples to oranges. But even with that taken into account going with a v4 makes more sense. They are within $10 of what the comparable v3 versions cost and for that additional amount you get a good increase in efficiency and reduction in temps+heat output and 5% higher IPC. The most apples to apples comparison would be the 2697v3 vs the 2690v4 as they both have 14 cores. The 2690v4 costs $25 and uses a bit less power (the TDP can be misleading, under full load the 2690v4 will top out at 125W whereas the 2697v3 will top out at its 145W power limit). A few bucks more gets you 5% more IPC, 20W less real world power use, and faster memory support. Seems like an easy choice really.
Also, why in the world would you disable Hyper-Threading to run CPU benchmarks? Most people are not going to do that and SMT being enabled on high core count CPUs has not been a problem in gaming for quite a few years now. Going back to the non-gaming benchmarks you ran, disabling SMT on the Xeons basically made the results irrelevant and invalid. SMT is enabled by default and increases performance in multi-threaded workloads by 20% on Intel. So in reality your scores on those should be about 20% higher than they are. Appreciate the time and effort, but the testing methodology could use some work.
The E5-2690 V4 is on its way to me, hopefully. The Hyper-Threading was disabled because it often yielded worse results in games, especially the 1% and .1% lows were significantly better without HT the last time I tested. Maybe it's time to re-test it and see what I can get with HT enabled but nobody is going to toggle Hyper-Threading between gaming and non-gaming tasks, thus I tested the configuration that I use myself in my Xeon gaming PC and that makes the most sense for gamers.
i got a b250 mbo and a 7600k, its getting pushed to 100% in the finals and the game is lagging because of it, what should i upgrade to? a xenon e5 of sorts with a x99 board is around 120 euro at least it seems like it, the qtj1 lga 1151 is also 120 euro, the qtj1 2.6ghz board is 200 euros or a ryzen 5 5500/5600 for 240 euros with ram? ryzen has an upgrade path... but then again next time i think il just buy full pc as i got a rtx 2060 only so...
boa tarde, eu procurei informação sobre o xeon e5 1660 v4 e não encontrei em nenhum site , este processador é bom para jogos? gostaria por favor de sua opinião. Um abraço
@@Miyconst will give it go on the MR9A PRO board from Machinist this weekend. should have a 280 aio somewhere (hopefully with compatible mounts). 4.5 ghz, maybe ;) ???
@@Miyconst received a mr9a pro board and lo behold the seller had a misleading listing. instead of q87 chipset received a b85 and ofcourse no overclocking for the i7 as planned. ali express is a minefield when it comes to seller reliability. now thinking of returning or just buying a 1660 v3
Neither will be equivalent Just buy the 6950x or 5960x get an AIO OC and higher ram speed get an X99 A motherboard. They both whip Xeon's in single performance.
Hi Miyconst, your videos are awesome. I'm a happy x99 platform user. My 2640v3 and my Huananzy 8mf performs very very well. I use it with a Radeon 6600 and I'm very happy with the performance in games, but I think I have a little bit of bottleneck in games because CPU and GPU rises normally between the 70-80%, sometimes the GPU is 100 and CPU 80 and sometimes the reverse, but most of the time is 70-80, 80-80, and numbers like those. Ram in 2640 v4 runs @ 1866mhz, what do you think to update the CPU to a 2667 v3 and get a bit more of mhz and a usable ram at 2133mhz (my ram is 2400mhz). It's only 20 euro, very affordable. If you think it's a good option, it's only necessary to swap the CPU's or due the turbo unlock I need to do something more? Thanks for your job!!!
If you are satisfied with the performance - there is no need to change anything. If you unsatisfied, you need to identify what part is slowing you down and upgrade it. E5-2666 V3 would be a good upgrade option and it also works with TBU BIOS. E5-2667 V3 is a possible alternative but since the price is the same as E5-2666 V3 I don't see why you should pick less cores for the same money.
Hello dear, I really appreciate your good work on your videos, you give us lots of valuable information about this platform and pc overall, i really want to buy an 2011-3 cpu today and I'm confused should i buy e5 2697v3 or e5 2667v4, I've watched you video about machinist mr9a v1 and i bought it, so i need just a cpu now, i want it for gaming. Thank you.
I think I have provided more than enough information about these two CPUs in my videos, if you have any particular questions - please join my Discord and ask there.
@Miyconst i didn't find an actual comparison between the two, I've seen your videos about each one but it's hard to decide with one to pick so i want you advice if you can. sorry for bothering you.
I got an older Processor Intel Core i5 4460 ... so i thinking buy a motherboard for that processor on aliexpress .. the Motherboard i thinking buy is that..(Huananzhi-placa-mãe h97 zd3 h97, lga 1150 m-atx, sata3, usb 3, nvme, ngff, m.2, ssd, não-ecc, núcleo i3 4130, i5, 4460, i7 4770) .... i think with this system i got less problems.. What you thinking ??
You should pair it with Nvidia cards to see what kind of difference there is. Is it possible to compare the best price to performance Xeon(s), to all the valid price to performance chips of current time? Tweaking and maximizing the limits, pushing it to areas probably never achieved on the Xeons? Thank you!
@@Miyconst I'm not 100% certain on this one (most likely a very silly one) but I think AMD GPUs work well with AMD CPUs, as maybe Intel and Nvidia might synergize more better? This is probably very false, but I don't think it hurts to try (which you may have already did because I just started watching your vids and have only seen the more recent ones). Thank you for the very informative videos!
I have one paired up with my 6900xt. The pair is a match made in heaven. Due to my 1080p monitor i can get up to 90% CPU usage. GPU is pretty much always pegged at 98%-100%. Can't wait to get a 1440p or 4k monitor in the future to take some work off the R5 5600x3d.
im just wondering tho how much it would make a difference if possible i would really like to see i like your vids btw its helping me build a 2697 v3 gaming computer @@Miyconst
Great content, as usual, very professional and "to the point". As an IT specialist I have access to plenty of server hardware: last year, just for fun, I purchased one of those chinese "kits" with a 12-core old Xeon. I have to say that this machine, with 64Gb ecc ram, performs wonderfully as a KVM/Virtualization server under Linux, as a Backup server and for developing sotware (pythin/php mostly). Running for 7 months now, can't be happier, the price was truly unbelievable (I guess I paid about 115€ for the whole thing), and the machine performs flawlessly, on 24/7 in the cold room. I think gaming software is totally the wrong use case for those old Xeons: they were built for virtualization/database/backup backbones, not really fit for something like a videogame. For that, a single-core beast (like a cheap i3 12100, or a 13100, AMD 5500 5600 ecc) is BY FAR a better choice for less than 20% more money. I am purchasing another one I've found on aliexpress with a 2680v4 and 32gig for 110€, it's really a ridicolous amount of computing power for such a price.
I just swapped out an i7-5930K for the 2697A v4 on my Gigabyte X99 board. The previous owner of the i7 I assume had been overclocking the poor thing to the moon cuz it was unstable with even a tiny bump. I went Xeon specifically because I'll know it hasn't been abused. So far, performance is basically the same in everything I do, but I'm happy with it. It's fun having 16 cores and it was $33. Has anyone tried disabling some cores on these high core count chips? I wonder if running it with 8 would still give me the full 40MB of cache? Games only really expect 8 cores... I already disabled HT. My BIOS supports setting the active cores to any number, maybe I'll try it tonight.
@@Miyconst I did some experimenting in Enshrouded which is CPU bottlenecked on my system. I compared 16 cores to 8, 6, and even 4 core with HT. Nothing improved FPS vs 16. 4 core with HT got slightly higher clock speeds and a few FPS higher when staring at the floor, but any action caused tons of stuttering. I'll leave it at 16 w/o HT for now. It was a fun experiment, but there are no poor man's Xeon X3D's to be had unfortunately. ;)
Another excellent video. Really the old Xeons don't make sense for desktop computers anymore. Cheap Ryzens win every time. And the cost on Skylake (and newer) Xeons are just way too expensive. I think the high core count still makes sense for a Proxmox virtualization server. If you need to allocate cores to a VM you're going to run out of Ryzen cores very quickly. Pretty much every other case by a Ryzen.
You can share CPU cores between VMs so less cores is not an issue as long as these are strong cores. Memory bandwidth is another story, quad channel is definitely handy when multiple VMs load the memory controller at the same time.
They make a ton of sense if you want cheap multi-threaded power, and VMs are not the only use case for that. It can be anything from video encoding, 3D rendering, code compilation, file compression/decompression, distributed computing, and other things. Building a PC using an older Xeon doesn't really make sense though... maybe it does if you get a dual CPU motherboard for running tasks that scale well with many-core CPUs. The cost is way, way higher than getting a barebones HP Z440 with a 700W PSU+heatsink for under $100 or similar and slapping 4x8GB or 4x16GB of DDR4 ECC RDIMMs and a $20-30 Xeon E5-2690 v4 14-core or 2682v4 16-core on it. There ain't no way in hell you're building a system that has that much multi-threaded performance for around $200 using a 2017 or newer CPU.
I'm building a machine and in the last second I changed from the 2696v3 to the 2697av4. Even if its slightly weaker in terms of raw performance, the newer architecture and supposedly better power efficiency make me feel its a better choice, especially since Ill be using a Chinese board, where VRMS are usually very weak.
A comparison between those two (with TB OFF / ON on the V3) would be great!
I absolutely love this content, I've made up quite a few Xeon (x79/x99) gaming rigs for people over the years and never had any complaints.
I've got a Xeon 2696v3 next to me that's going in a Dell 5810 just for fun.
Its mental that these things just won't die. You can pick up an x99 Xeon + mobo for around 60% the price of an 5600+mobo second hand so the price to performance is still there.
Very true. When the price is right, these Xeons are still very attractive.
I appreciate your channel. As an IT guy its fun to see how equipment I have dealt with for years runs modern games compared to newer stuff.
Great breakdown! A lot of your videos remind me of a few anantech articles that really broke the lga 2011 v3 cpus down in incredible detail. I remember they all remarked on how the v4 series pretty much just held the line at best and often reduced the avx clocks a little on all core loads. The ring bus core count also went up from 8 to 10 to compensate, however.
Looks like you were hitting a pretty solid 3.4-3.5 ghz avx with the v3
I bit the bullet and decided to go for a e5-2689 v4 and, while the records are incredibly in complete, I estimate it's all core avx turbo to be around 3.5-3.6. Fingers crossed. =)
Remember to keep us updated when that E5-2689 V4 gets tested!
@@Miyconst You bet! =D
@@beansnrice321 hey! how'd it went?
@@mxsugz Kind of oddly. It's faster in single threaded tasks but only barely. The 2690 seemed to do more with a lower tdp. I suspect that the memory controller on the 2690 is a little more substantial. I base this theory off of aida64 latency tests showing the 2090's higher cache was also faster and had lower latency. Main memory access was also slightly faster on the 2090.
Games like cyber Punk seem to run worse and hotter on the 2689 over the 2690.
However, however. I got it for one game and that game was Elden Ring and it seemed to really help exactly where Elden Ring needed it, which was with single core speed.
The main thread of Elden Ring really craps the bed if it maxes out the core it's running on. The 2689 gave me my 8% extra margin I needed to hit that limit just that much less often to make the game noticeably more fluid.
So I'm happy with the chip. I'm happier still with the price I paid, $60 US and I'm still happy to have the old 2090 to fall back on in case I do something like start doing more video production.
Fantastic content, with increasing support of multithreaded games and apps this xeons are holding pretty well, I was expecting far worse.
excellent work my friend!! It looks like your video came at the right time!
And all of us know very well that the best old Xeon CPU is the 2696v3! You can reach 3.6-3.8 GHz in all cores with 12/24 cores enabled!
That's very true! It also has massive amount of cache.
@@Miyconst But it costs a little more than other options. I still prefer it more than the V4 by now.
Great video as always, man
Are you watching me?! Lol. I'm literally sitting between a Xeon e5-2697a system and a R5 5600 system. My hope is to have one Proxmox server for VMs and primary storage and an Unraid server for Docker and backup storage. Love your informative videos, in particular in-depth looks at the X99 platform in modern day scenarios.
If you can afford to buy a 4 to 6 memory slot motherboard for ryzen 5 then it's good. But i think you will end up spending 5 times the price of the v4
😂, take care for your cat, thanks a lot for your benchmark
you're making great content!
Im leaving my experience here:
I have mutliple xeons in my hand and I concluded that the best xeon for gaming is the Xeon 1660 v3. I'll mention all the points of all the Xeons I have:
1: Xeon 2690 v3 - Great cpu for the price ($9). It can mantain full frequency after turbo unlock. For games like warzone struggles.
2: Xeon E5 2696 v3 - Best Efficiency / Perfomance you can have under x99. At some points it can compete with the Xeon 1660v3 (4.3ghz). However it uses like 60-90 watts less. Recommended to undervolt and disable 8-10 cores for 3.8ghz (unlock). If you are worried of watts go with this one. Due to high amount of cache memory and freq this can handle games as warzone (110-120 fps)
3: Xeon 1660 v3 - if you pick this one you don't care about your VRM/Comsuptions or any risks involved. It's inferior to other cheap xeons when used on stock. To go around 4.5ghz / 4.4 ghz you will have to go above 1.26v in vcore. Going below 4.3ghz cannot compete with xeon 2696 v3. Since this processor only has 20mb cache. However you can overclock cache to 3.5ghz and beat xeon 2696 v3. At this point your cpu/vrm/mobo would be dealing with ton of power.
With my Rx 6700 xt and Xeon 1660 v3 at 4.4/4.5 ghz I'm consuming around 460 watts! (Gpu undervolted even). Which means cpu is doing this lifting.
Gpu is taking 120-140 watts
Good job!! Thank you bro
Дякую ща вашу працю!
Great video..thanx on comparison, latley you like to turn hyperthreading off, better for gaming or you do it for another reason?
The main reason why I disable HT is because from my testing (conduced quite a while ago) the 1% and .1% low results are much better without it. An additional reason - I don't want to show that many threads in MSI Afterburner overlay. :D
Excellent work.
I think one of the most interesting results to come out of this video is the study involving the 16c/16t vs 8c/16t. I may be mistaken but the only real difference I see between the shutting of hyper-threading with the one test and shutting off 8 cores with the other is the cutting in half the cache resulting in a 10% increase in frame rate with the 16c/16t test relative to the 8c/16t test.
The L3 cache is shared among the cores so disabling cores doesn't affect the cache size.
Which choice would deliver more FPS, an e5 1660 V3 with overclock or an e5 2697 V3 with turbo Boost?
I would go with E5-2697 V3.
Very good. Thanks from brazil.
I have a 2697 v3 coming soon, my BIOS is patched with Ultimate patcher tool (x99 - 8M) and I was wondering how the turbo boost unlock works with it, and how to properly apply a voltage offset, thanks for the informative videos
Hi. How did it work?
@@Pablo-jk2lo Focusing on exams right now, so didn't have a chance to try, will update when I do
Dear Miyconst I am folowing yor channel long time now and I have one qoestion for you if you can reply, which one xeon for x79 or x99 is best to pair with gtx 1080 ti? Thank you :)
The simple answer is: i3-12100F, the complicated answer is E5-2697 V3, even more complicated answer: wait a bit until my set of V4 Xeons arrive and I test them, maybe there is a better option than E5-2697 V3.
@@Miyconst Thank you for reply, I really appreciate and I will wait for that :)
Thank you very as always very informative
I have big question i hope you'll answer me
I'd rather go with your favorite setup
E5 2697a v3 + 32g ram + huananzshi x99 tf
Or I'll go with
E5 2680 v4 + 32 ram + mr9a pro
Why not E5-2697 V3 with MR9A Pro? The X99-TF is overpriced and is not worth the extra price.
The mr9a is it the best go for the Xeon cpu? Is there another one?
super enlightening! Thank you for all the hard work! It's safe to say if you're building a new system you are way better off going with a ryzen 5 2600 and a b450m. Heck you can probably build a ryzen 5 3600 system for the same cost as the top of the line xeon x99 system, but the ryzen system will have way more room for upgrades in the future.
With Ryzen you not only get much better upgrade path but also much better motherboards.
Thats a shame about your camera set up but you have an awesome fluffy cat!
What is his/her name?
I think I've shared this before but my cat "Woolley" looks very similar to yours and happens to sleep belly upwards like shown in here in your video. Its a shame RUclips doesn't allow pics posted in the comments section.
He is called Zion and he is a Norwegian Forest Cat, the breed is similar to Maine Coon but not exactly the same.
Great video, do you think a 2697A will continue to bottleneck at 3440 x 1440 a 7800xt that much? I do have a QD4 huananzhi will it support the processor?
It really depends on what games you play. In games like CS2 the CPU will be the bottleneck. X99-QD4 should be fine with E5-2697A V4.
@@Miyconst Warzone for example? Dayz and shooters mainly
@@futboltop-ku1shyes, you most likely will be CPU limited.
Hello! Very interesting video! I have a Intel Xeon E5-2697A v4 with a RX 6500 XT and the performance are good. My problem is that I have a serious issue with a server motherboard Supermicro X10SRL-F, it's not suited for gaming and it burned my videocard. Have you any experience with server motherboard? What do you suggest for replacing it considering that I have 8x16GB DDR4 2400 ECC RAM? Thanks in advance!
Hi. If your motherboard fried your GPU then it's simply a defective unit, it has absolutely nothing to do with "gaming" or "server" motherboard. 8 x 16 GB is a total overkill for gaming, thus I am not sure why you need it. My go-to Chinese X99 motherboard is Machinist X99-MR9A (Pro).
@@Miyconst thanks a lot for the fast response. First reason is because I had it for free, second reason is because sometimes I virtualize devices and I need a lot of RAM. Can you suggest me a model with 4 ddr4 ecc slot? Thanks in advance for your content that's exactly was struggling in these months and I finally found someone with the information that I needed. I was very skeptical at the beginning with this Chinese motherboards. Keep up the good work!
@@federicopullini6206 I have already suggested you Machinist X99-MR9A (Pro) or it doesn't work for you?
@@Miyconst sorry I meant 8 slot
@@federicopullini6206 Huananzhi X99-F8 is the only one.
how do you feel about the e5-2697 v4 vs the e5-2697a v4 debate.... i currently have the e5-2697 v4 and have contemplated getting a e5-2697a because of the higher base clock which is .2 ghz more which im thinking is not going make much of a difference. id be intrested at seeing a side by side comparison of the two all the videos i find are not in english etc
The performance difference will be negligible.
Excellent video ... But I have a huge doubt for my case friend ... I have a Huananzhi QD4 Motherboard with 32 gigs of ECC quad ram .. and I have a Xeon 2697A V4 processor currently in use on this motherboard ... but when I bought the KIt it came with a Xeon 2670 V3 Processor ... I made a mistake by putting a stronger processor ... since the one I'm using has 145 TDP and the other 120 TDP ... will I damage the motherboard in the long run ?? is that when the processor pulls I notice that the VRM heats up a lot ... with Xeon 2697A V4, but it also heats up with the Xeon 2670 V3 ... after all, where do we stand ?? I would like to hear an opinion. ??
With either of these CPUs you must have good cooling and good airflow inside your chassis. If the VRM heats up with E5-2670 V3, I can assume that you have very bad cooling inside the chassis or at least the VRM is not properly cooled. So with E5-2697 V4 you are really risking to burn it down.
@@Miyconst Thank you... for your opinion .. so in this case it's better the processor Xeon 2670 V3.. ? because the TDP is more Lower? .. actually the temperature of processor is around 32º degrees celsius of course with ( Xeon 2697A V4).. what you think?
Can you use quickcpu tool to scale max mhz on e5 2697a v4 and test it
Watching the Alan Wake DX12 benchmark-- for some reason the V4 keeps constantly downclocking its cores to 300mhz while the V3 does not 🧐 that kind of throttling might raise 0.1% lows if not introduce microstutters and/or affect the whole run.
This is not throttling. The CPU cores drop the clock when there is no load on that core or when the scheduler decides to switch cores that do the job.
The only v4 CPU worth considering at the moment is the E5-2689 v4, which is basically an overclocked and undervolted i7-6950X.
It can be hard to find, and depending on its price, you might as well just go with an i7-6900K or 6950X, since they retain their value much better and for longer due to the "i7" branding.
Yeah, the E5-2689 V4 is very interesting but the price is bad and it's not many of them in the wild.
Hey guys, do I get an e5 2690v3 or an e5 2680v4? both same price so which is better? v3 for the overclocking or v4 because its newer
Hello Miyconst, On BIOS I have option for disabling cores, for E5-2697A v4 (or in general for XEON e5 cores) do you think it makes any sense? It can improve performance or reduce significantly the power consumption? Or it's just not relevant?
Thanks!
It really depends on the CPU you have and your use-case. If you are not sure - try it and see what are the results.
Would it be good idea to pair 7900xtx or 4070 ti Super with 2697a v4, 2697 v4, 2699 v3 ?
I'm concerned regarding bottleneck.
My motherboard is Msi x99A plus, support up to 2699 v4, which is much more expensive than v3.
Which combination would you recommend?
I'd like to play 4K with 70-120fps.
Thank you for your great videos
With the MSI board you can get i7-6950X and overclock it or just sell it all together and get something modern.
@@Miyconst I'd prefer to slightly upgrade it as cpu should be capable yet to support most of top end graphic cards.
I currently have i7 5820k
@@Miyconst Thank you,. didn't know this processor would be the best for gaming for my system.
@@Miyconst does it worth upgrading from 5820k to 6950x ?
Or the results will be similar?
Great content regarding e5-2600v4. I have an e5-2680v4 (14c,28t) CPU and my R23 score is 12127. Strange that your score is so low in comparison. Could it be the Chinese motherboard? I am using an Asus X99 Prime A-ll with 96GB of ECC DDR4. My system has been rock solid since I cobbled it together a year ago. Most stable system I have ever put together. it's used for scientific computational modelling.
I assume you have Hyper Threading enabled with your E5-2680 V4, right? In that case it explains why your score is higher, you got 14 Cores / 28 Threads, I tested 16 Cores / 16 Threads.
Yes, you are correct. My bad. @@Miyconst
Nice
Im still enjoying my E5-2697v3. On my Rampage V extreme motherboard. Recently got a RTX 3070 so need to watch and do the Re-bar mod.
I need a suggestion. Currently im runing 12c/24t and the cpu goes to 3.5ghz, any baseclock oc makes the system show no picture with the RTX 3070. (I used a Titan X maxwell woth 105mhz baseclock and the cpu went to 3.7ghz with 12c)
What should i do? Keep the cpu 12c/24t at 3.5ghz
14c/14t or go 10c/20t? (10c active the boostspeed goes to 3.6ghz)
I would just keep the 12c/24t configuration, seems like the sweet spot. Resizable bar makes little to no difference with Nvidia GPUs, so I wouldn't bother.
@@Miyconst Alright noted and thank you for your help :)
Hi I need your opinion on 1680v3 vs 1650v4 gaming performance which you think is a better choice i want to build it on hp or dell workstation
I think E5-1680 V3 is gonna be better.
uma placa mãe machist PR9 aguenta esse processador sem unlock? estudo, trabalho e jogos
What CPU are you asking about? I tested two 😀.
xeon 2697 V3 @@Miyconst
Hello friend, the e7 8800 v4 is compatible with an x99 motherboard.
E7 CPUs use another socket, even though its called "LGA 2011-3", it's physically different, Intel really did "good job" naming these sockets.
"E7-8890 v4 is LGA2011-1 which is not compatible with X79 (LGA2011) nor X99 (LGA2011-3) boards.
LGA2011-1 was only used in high-end servers with up to 8 sockets"
Miconst how does a i7 7820x compare with the 2667v4, ive seen a seller in my area offering a dell t5820 with i7 7820x about 160$, and the same guy is also listing a hp z440/dell t5810 with xeon 2667v4 for about 120$, is the newer platform better in terms of gaming, because i have seen reviews of skylake x cpus running rather hot and inefficient and also do they performance rather poorer in games, I have no way to check as I only have an old laptop, and I don't want want to buy new stuff, it's out of range for me, humbly awaiting your response
That's a very tough choice to be honest. The LGA 2066 CPUs are indeed very inefficient and bad for gaming but you get faster memory and an option to upgrade to other i7/i9. If it would be me, I would probably pick T5810 as it's cheaper and invest the savings into a better SSD/GPU/monitor.
I use the E5-2697A V4 + RX 6700XT on a Huananzhi F8 mainboaard. A set of cheap HyperX DDR3200 non ECC ram 32 GB. Granted I only play Call of Duty, it actually plays it pretty ok. I have HT enabled. I have installed the Rebar unlock but it causes some instability so have it turned off in the drivers. The 2697A V4 is a bit of a rare CPU so probably a bit more expensive compared to other E5 counterparts. Call of Duty is known to scale well with a lot of cores and likes memory bandwidth so no surprise there that it does well. I get a slightly higher score in R23 though, roughly 14.000 not sure what is different there (maybe because I use HT?)
This E5-2697A system is a few years old now though, for the low cost / performance segment I would not turn to the Xeon anymore. This will likely continue as a local NAS/Visualisation server for me though and there it does the job fine.
The Cinebench score difference is surely because of HT, I disabled it.
Call of Duty plays with these old Xeons surprisingly well, I can confirm that.
Hi, can you do a review on RX Vega HBCC over Xeon Quad Channel RAM setup ? It can sove the 8GB VRAM issue that some games stutter upon nowadays.
Vega 64 is not strong enough to have any meaningful use more than 8 GB vRAM. There are already quite a few videos on this subject and the quad channel memory configuration won't change the overall picture.
@@Miyconst My attempts to see HBCC on a quad channel setup failed so far, but I take your word on that. I have such a setup at home and the memory bandwidth is huge. Most of the HBCC videos I see are limited by wimpy chipsets and narrow memory bandwidth, so I was looking for something in this specific combination. Many thanks again for the answer and Slava Ukraiini!
@@Miyconst sorry for the necro post. I've tested it and got a nice bump on Quad Channel DDR3 for E5 2680 V2. In 1440p with RX5700 I could not get more than 28 FPS in Talos Principle 2 using FSR3, with mildly overclocked RX Vega 64 I got 40+. The HBCC did its job and met some speedy memory there. VRAM usage jumped from 8GB to 10+ and the extra RAM bandwidth was observed at dynamic texture loading. On dual channel DDR4 3000 it was sensibly slower than Quad Channel DDR3 1866. Not to mention that for the price of 16 GB desktop DDR4, one can get 64 GB server DDR3.
What e5 cpu is best for gaming ? Now i have a 6800k. Will upgrade to e5 be better for me ?
It depends on the games you play and how far your i7 is overclocked. In general my advice is to move to AM4 or LGA 1700.
@@Miyconst but i dont have money for buy a new settup. Just wonder if a E5 i better. I have gpu 6950xt.
@@cripplefpv7354I can't add more, whether Xeon E5-2697 V3 is better depends on your i7's overclocking level and what games you play, also whether you apply or don't apply TBU.
How many BPS can you cat achieve
BPS = biscuits per second
I need to buy a BPS-meter to verify :D.
Hi Is it possible to overclock my xeon 1660 v3 with machinist mr9a x99?
or buy xeon 2697 v3
Theoretically it's possible but so far I am not aware of a BIOS that supports overclocking on MR9A, so I would pick E5-2697 V3 instead.
@@Miyconst alr ty
@@robinbambach1458 i bought 1080 ti for 145 dollars
Comparing a turbo unlocked v3 to stock v4 is kinda apples to oranges. But even with that taken into account going with a v4 makes more sense. They are within $10 of what the comparable v3 versions cost and for that additional amount you get a good increase in efficiency and reduction in temps+heat output and 5% higher IPC.
The most apples to apples comparison would be the 2697v3 vs the 2690v4 as they both have 14 cores. The 2690v4 costs $25 and uses a bit less power (the TDP can be misleading, under full load the 2690v4 will top out at 125W whereas the 2697v3 will top out at its 145W power limit).
A few bucks more gets you 5% more IPC, 20W less real world power use, and faster memory support. Seems like an easy choice really.
Also, why in the world would you disable Hyper-Threading to run CPU benchmarks? Most people are not going to do that and SMT being enabled on high core count CPUs has not been a problem in gaming for quite a few years now.
Going back to the non-gaming benchmarks you ran, disabling SMT on the Xeons basically made the results irrelevant and invalid. SMT is enabled by default and increases performance in multi-threaded workloads by 20% on Intel. So in reality your scores on those should be about 20% higher than they are.
Appreciate the time and effort, but the testing methodology could use some work.
The E5-2690 V4 is on its way to me, hopefully. The Hyper-Threading was disabled because it often yielded worse results in games, especially the 1% and .1% lows were significantly better without HT the last time I tested. Maybe it's time to re-test it and see what I can get with HT enabled but nobody is going to toggle Hyper-Threading between gaming and non-gaming tasks, thus I tested the configuration that I use myself in my Xeon gaming PC and that makes the most sense for gamers.
What should you take the E5-2697 V3 or v 4 =?
It really depends on your priorities. Ultra low budget - V3. Better power efficiency and no BIOS mods - V4.
i got a b250 mbo and a 7600k, its getting pushed to 100% in the finals and the game is lagging because of it, what should i upgrade to? a xenon e5 of sorts with a x99 board is around 120 euro at least it seems like it, the qtj1 lga 1151 is also 120 euro, the qtj1 2.6ghz board is 200 euros or a ryzen 5 5500/5600 for 240 euros with ram? ryzen has an upgrade path... but then again next time i think il just buy full pc as i got a rtx 2060 only so...
Please do yourself a favor and move to AM5/LGA 1700 or AM5.
boa tarde, eu procurei informação sobre o xeon e5 1660 v4 e não encontrei em nenhum site , este processador é bom para jogos? gostaria por favor de sua opinião. Um abraço
Xeon E5-1660 V4 is locked, not possible to overclock. It's ok for gaming if the price is right but nothing exceptionally good about it.
@@Miyconst Bom dia, muito obrigado pela atenção e um abraço
friend the E5 2673 v3 vs e5 2697a v4 which is better for games and tasks
I would go with E5-2697 V4.
could you test the 2680v4 versus the other two xeons you used in this video?
That's what I am trying to do right now..
Wouldn't 2696v3 @ 3.8GHz (HT off, maybe some cores off) be best option for gaming?
Yes, E5-2696 V3 is indeed the best gaming X99 Xeon out there (if we skip the unlocked and overpriced parts).
i know the 5960x is not the best pairing with a chinese x99 (q87) board but have one lying about. should sit somewhere between the v3 and 5600?
It really depends on how far you can push overclocking, in stock it's pretty pathetic.
@@Miyconst will give it go on the MR9A PRO board from Machinist this weekend. should have a 280 aio somewhere (hopefully with compatible mounts). 4.5 ghz, maybe ;) ???
@@Miyconst received a mr9a pro board and lo behold the seller had a misleading listing. instead of q87 chipset received a b85 and ofcourse no overclocking for the i7 as planned. ali express is a minefield when it comes to seller reliability. now thinking of returning or just buying a 1660 v3
1660 v3 or 1680 v3 which is equal to i7 6950x ?
Neither, the i7 is 14nm, the V3 as 22nm.
@@Miyconst which one is equal to 6950x ?
Neither will be equivalent Just buy the 6950x or 5960x get an AIO OC and higher ram speed get an X99 A motherboard. They both whip Xeon's in single performance.
@@RealTechGeek x99 A from AliExpress ?
Not here for your beautiful face but I must say these glasses look good on you!
Thanks :)
Hi Miyconst, your videos are awesome. I'm a happy x99 platform user. My 2640v3 and my Huananzy 8mf performs very very well. I use it with a Radeon 6600 and I'm very happy with the performance in games, but I think I have a little bit of bottleneck in games because CPU and GPU rises normally between the 70-80%, sometimes the GPU is 100 and CPU 80 and sometimes the reverse, but most of the time is 70-80, 80-80, and numbers like those. Ram in 2640 v4 runs @ 1866mhz, what do you think to update the CPU to a 2667 v3 and get a bit more of mhz and a usable ram at 2133mhz (my ram is 2400mhz). It's only 20 euro, very affordable. If you think it's a good option, it's only necessary to swap the CPU's or due the turbo unlock I need to do something more? Thanks for your job!!!
If you are satisfied with the performance - there is no need to change anything. If you unsatisfied, you need to identify what part is slowing you down and upgrade it. E5-2666 V3 would be a good upgrade option and it also works with TBU BIOS. E5-2667 V3 is a possible alternative but since the price is the same as E5-2666 V3 I don't see why you should pick less cores for the same money.
@@Miyconst will check both, thanks a lot :)
@@Miyconst finally I get a 2666v3 for only 13eur, I think it's a good update 😁
Can a MR9A v1 support a 2697 v3? Thx.
Yes, it can.
отчего тогда такая неадекватная цена на 2697ав4, он же абсолютно безхайповый?!
how about Xeon E5 2697a V4 vs Xeon E5-2699A V4
E5-2699A V4 has massive 55 MB of L3 cache and 3.6 GHz max frequency so with a few cores disabled it might be good for gaming if the price is right.
Hello dear,
I really appreciate your good work on your videos, you give us lots of valuable information about this platform and pc overall, i really want to buy an 2011-3 cpu today and I'm confused should i buy e5 2697v3 or e5 2667v4, I've watched you video about machinist mr9a v1 and i bought it, so i need just a cpu now, i want it for gaming.
Thank you.
I think I have provided more than enough information about these two CPUs in my videos, if you have any particular questions - please join my Discord and ask there.
@Miyconst i didn't find an actual comparison between the two, I've seen your videos about each one but it's hard to decide with one to pick so i want you advice if you can. sorry for bothering you.
@@seraph.919 the conclusion is obvious - V4 Xeons without turbo boost unlock are pointless. Buy e5 2697 v3.
I got an older Processor Intel Core i5 4460 ... so i thinking buy a motherboard for that processor on aliexpress .. the Motherboard i thinking buy is that..(Huananzhi-placa-mãe h97 zd3 h97, lga 1150 m-atx, sata3, usb 3, nvme, ngff, m.2, ssd, não-ecc, núcleo i3 4130, i5, 4460, i7 4770) .... i think with this system i got less problems.. What you thinking ??
I think it's a waste of money.
@@Miyconst ok … and about professor do i have changed the Xeon e5 2697a v4 to Xeon e5 2670 v4.. because the TDP ismore lower??
@@icecrypt124 Because its frequency is lower, then a worse option for gaming.
You should pair it with Nvidia cards to see what kind of difference there is. Is it possible to compare the best price to performance Xeon(s), to all the valid price to performance chips of current time? Tweaking and maximizing the limits, pushing it to areas probably never achieved on the Xeons? Thank you!
Maybe I don't understand something but what would be the difference with an Nvidia GPU?
@@Miyconst I'm not 100% certain on this one (most likely a very silly one) but I think AMD GPUs work well with AMD CPUs, as maybe Intel and Nvidia might synergize more better? This is probably very false, but I don't think it hurts to try (which you may have already did because I just started watching your vids and have only seen the more recent ones). Thank you for the very informative videos!
I guess you didn't see this video: ruclips.net/video/TZ5T2aor9pI/видео.htmlsi=0SPSf_k82EoMKh_9
@@Miyconst Very informative as always! Thank you once again!
E5-2695 v4 18 core is what I settled on after trying about 5 or 6.
You definitely should get the Ryzen 5 5600X3D. It's an mini beast!
They don't sell them in Sweden..
@@Miyconst Yeah I know it's an microcenter exclusive. We all gotta figure out an way to get you one.
I have one paired up with my 6900xt. The pair is a match made in heaven. Due to my 1080p monitor i can get up to 90% CPU usage. GPU is pretty much always pegged at 98%-100%. Can't wait to get a 1440p or 4k monitor in the future to take some work off the R5 5600x3d.
how about 2697 v3 with various rams test
with speed 2133 vs 2666
Max possible RAM speed with E5-2697 V3 is DDR4-2133.
@@Miyconst just call it OC possible?
@@buzzbang9164 it's not possible.
@@Miyconst you just did it bro 😅
If the ram speeds were the same im guessing the ryzen would be much more level with the v3 (however i see that's part of the setback)
Well.. we could say the same about the clock frequency as well. Xeon is locked and doesn't support higher RAM or CPU clocks, so it is what it is.
im just wondering tho how much it would make a difference if possible i would really like to see i like your vids btw its helping me build a 2697 v3 gaming computer @@Miyconst
@@GoDDooMzJulz8145there is no possibility to make scientific measures of such a CPU and I am not going to speculate, it won't help anybody.
Great content, as usual, very professional and "to the point". As an IT specialist I have access to plenty of server hardware: last year, just for fun, I purchased one of those chinese "kits" with a 12-core old Xeon. I have to say that this machine, with 64Gb ecc ram, performs wonderfully as a KVM/Virtualization server under Linux, as a Backup server and for developing sotware (pythin/php mostly). Running for 7 months now, can't be happier, the price was truly unbelievable (I guess I paid about 115€ for the whole thing), and the machine performs flawlessly, on 24/7 in the cold room. I think gaming software is totally the wrong use case for those old Xeons: they were built for virtualization/database/backup backbones, not really fit for something like a videogame. For that, a single-core beast (like a cheap i3 12100, or a 13100, AMD 5500 5600 ecc) is BY FAR a better choice for less than 20% more money. I am purchasing another one I've found on aliexpress with a 2680v4 and 32gig for 110€, it's really a ridicolous amount of computing power for such a price.
I just swapped out an i7-5930K for the 2697A v4 on my Gigabyte X99 board. The previous owner of the i7 I assume had been overclocking the poor thing to the moon cuz it was unstable with even a tiny bump. I went Xeon specifically because I'll know it hasn't been abused. So far, performance is basically the same in everything I do, but I'm happy with it. It's fun having 16 cores and it was $33.
Has anyone tried disabling some cores on these high core count chips? I wonder if running it with 8 would still give me the full 40MB of cache? Games only really expect 8 cores... I already disabled HT. My BIOS supports setting the active cores to any number, maybe I'll try it tonight.
Just try and run some tests. The L3 cache is shared among all the cores so disabling a few doesn't reduce the amount of cache.
@@Miyconst I did some experimenting in Enshrouded which is CPU bottlenecked on my system. I compared 16 cores to 8, 6, and even 4 core with HT. Nothing improved FPS vs 16. 4 core with HT got slightly higher clock speeds and a few FPS higher when staring at the floor, but any action caused tons of stuttering. I'll leave it at 16 w/o HT for now. It was a fun experiment, but there are no poor man's Xeon X3D's to be had unfortunately. ;)
0:28
who knows buddy ,may be on this you've been falsely assuming so ,so happy new year and keep it up ...
Happy upcoming New Year to you as well. 🎇
Boa tarde, um abraço
Another excellent video. Really the old Xeons don't make sense for desktop computers anymore. Cheap Ryzens win every time. And the cost on Skylake (and newer) Xeons are just way too expensive. I think the high core count still makes sense for a Proxmox virtualization server. If you need to allocate cores to a VM you're going to run out of Ryzen cores very quickly. Pretty much every other case by a Ryzen.
You can share CPU cores between VMs so less cores is not an issue as long as these are strong cores. Memory bandwidth is another story, quad channel is definitely handy when multiple VMs load the memory controller at the same time.
They make a ton of sense if you want cheap multi-threaded power, and VMs are not the only use case for that. It can be anything from video encoding, 3D rendering, code compilation, file compression/decompression, distributed computing, and other things.
Building a PC using an older Xeon doesn't really make sense though... maybe it does if you get a dual CPU motherboard for running tasks that scale well with many-core CPUs. The cost is way, way higher than getting a barebones HP Z440 with a 700W PSU+heatsink for under $100 or similar and slapping 4x8GB or 4x16GB of DDR4 ECC RDIMMs and a $20-30 Xeon E5-2690 v4 14-core or 2682v4 16-core on it. There ain't no way in hell you're building a system that has that much multi-threaded performance for around $200 using a 2017 or newer CPU.
im here both for the benchmarks and ur beautiful face, 😄
Bad fluffy cat 😔😭
2697é o melhor acabei de montar um 2697 v3 + quyida qd4 x99
my e5 2697Av4 scores 1000 points more in cpuz 3:14
Probably you have HT enabled.
@@Miyconst yes it is enabled
ta trancando os processador ridiculo o 97A pega mais que isso