Is "AMDip" Real? - AMD vs Intel CPU Consistency
HTML-код
- Опубликовано: 5 сен 2024
- Is “AMDip” Real? - AMD vs Intel CPU Consistency
I’ve been seeing more and more comments about a supposed AMDip that AMD Ryzen CPUs have due to the distance and latency caused by the infinity fabric interconnect that bridges CCD CPU clusters and IOD silicon in AMD CPUs such as the Ryzen 7 7800X3D and Ryzen 9 7950X3D. Today we compare a 7800 X3D to an Intel 12900K to see if these claims have any merit.
Get access to the discord: / thedisplayguy
FULL DISCLOSURE
Amazon has given me affiliate links which send me a percentage of each purchase when you use them. Additionally all Vip-cdkdeals links are sponsored.
END DISCLOSURE
Specs
Camera: Canon EOS R6: amzn.to/3V2RnnL
RAM: 64GB Gskill DDR5: amzn.to/3F1Pa6j
CPU: Intel i9-13900K: amzn.to/3EaqGpD
MOBO: EVGA Z690 Kingpin: www.evga.com/p...
GPU: AMD RX 7900 XTX Aqua: click.linksyne...
Storage: Kingston 4TB Renegade SSD: amzn.to/3OXvdlz
Case: Lian Li O11 EVO: amzn.to/3VUscF3
Thanks to Vip-cdkdeals for sponsoring this video. 30% off code: GPC20
▬ Windows 10 pro ($17):www.vip-cdkdea...
▬ Windows 11 pro($23):www.vip-cdkdea...
▬ Windows 10 home ($14):www.vip-cdkdea...
▬ Office 2016($28):www.vip-cdkdea...
▬ Office 2019($47):www.vip-cdkdea...
Get access to the discord and ICC profiles: / thedisplayguy
Music
www.davidcutte...
Outro: Everyone by Monplaisir
Is “AMDip” Real? - AMD vs Intel CPU Consistency
I’ve been seeing more and more comments about a supposed AMDip that AMD Ryzen CPUs have due to the distance and latency caused by the infinity fabric interconnect that bridges CCD CPU clusters and IOD silicon in AMD CPUs such as the Ryzen 7 7800X3D and Ryzen 9 7950X3D. Today we compare a 7800 X3D to an Intel 12900K to see if these claims have any merit.
Get access to the discord: patreon.com/TheDisplayGuy
Buildzoid’s safe A-die timings work for both 6400MT/s CL30 and 6000MT/s CL28. He has tested his timings extensively and I have encountered zero issues or instability. I have encountered no performance drops as both memory configs are operating at 1:1 ratio. The sweet spot seems to be 2167 for the FLCK for my 7800x3d. I can go as high as 2200 but this is where I encounter “dips” and drops in FPS. The problem with tuning Ryzen 7000 is that it’s more difficult than Intel and more time consuming. Retraining the RAM while testing sub timings can be frustrating at best. I have used both XMP and EXPO memory. I tend to buy the highest clocked A-die and detune without using the preconfigured profiles. If I wanted to play it safe I would follow your advice but I want the most performance I can get that is stable, so I tune my system myself. I will say tuning Intel is a lot easier and I guess seeing your RAM operate at 7200MT/s, 7600MT/s, and 8000MT/s and benching higher in AIDA64 and Super PI is more exciting than being stuck at 6000MT/s and 6400MT/s but, that is all. I tune and use both Intel and AMD systems and the dips are usually result of poor memory and CPU voltage/frequency configurations. Hardware such as Motherboard and even your operating system and drivers can impact stability and performance. I wish this video was a little more in depth as again we are just seeing bar graphs.
Edit: running everything in “Auto” and not tuning either Intel or AMD systems leaves a lot of performance off the table.
Frame Chasers going mainstream 💀
Frame Chasers was also using W10 not W11. Which Ryzen 7000 is fully intended for W11. W11 has better CPU instructions for the AM5 platform that W10 just did not get. I really wish Frame Chasers would retest with W11. Which is wild to me that he is using W10 when Intel has straight up said you're supposed to be on W11 for the P/E cores to work as intended on the 12th/13th/14th gen Intel chips.
FC sells products. His video content is intended to make his products look good all while maintaining this facade of "honesty"
I spent some time with Jufes at CES and he is a great guy … super sharp.
@@chiron03
All im sayin... is...
FUCK!
@@chiron03 W10 is optimal for any Zen CPU, W11 is doing something weird with L3 cache
Yes, I can confirm the AMDip, based on the Intel (i7 13700k) and AMD (R7 7700x) builds I have. You dont see the DIP by showing benchmark numbers.
Just grab a very memory intensive game and move the mouse in an open scenario and you see / feel the DIP.
Very weak analysis...
Trust me bro ke
@@MrEdioss But it's true. If you start moving the mouse around especially if it's a high refresh mouse you will see the frame rate dip on the AMD CPU. Happens to me on a 5800X3D. It doesn't really bother me though because who's moving their mouse around really fast all the time in game. All this guy done was run a benchmark and comparing 0.1% lows and that doesn't cut it for testing the AMDip. He should do the video again and actually move the mouse around during his benchmark and see what happens then 😉
Do the same on Intel CPU and it's doing the same, it's dipping
@@RmX. AMD = much higher dips. Also, it has worse idle and light load power consumption. That's why I am currently just using the Intel setup for work + gaming. R7 7700x is dogshit for productivity.
@@paulotakahama Wasn't that a fTPM issue?
the amdip happens on 2 ccd cpus when the thread is moved to the ccd that doesnt have the data in the cache
Yeah but it’s easily solved with process lasso to set game affinity to only one ccd. Using process lasso the 2ccd layout is more convenient because I can use software encoding in obs on one ccd and run the game on the other with no real hit to my performance
@@jtake9896 true but what if you have 32 thread app and you want to keep the threads more or less on the same core ? i am not sure how the task scheduler will optimize execution. would be nice to hear if anything was implemented to optimize this.
@@builtofire1 if you keep affinity to one ccd, it will not use 32 threads for the app unless there’s an override that allows the app to ignore user set affinities
Frame Chasers says whatever will sell his products and services. Dude is a clown.
This comment usually goes both ways before we can get into the details. So far we haven't, and this video didn't prove anything either.
Hmmm, well back on my old Ryzen 1800x I DID have these weird dips, like a hiccup almost, it really only noticeable in VR though, but it being in VR it was also really uncomfortable.
U said it youre self back in the day thats the same as telling that amd drivers suck because they did back in the day now they are so good
The AMDip problem comes from the FCLK & CCD & IOD voltages. When you push FCLK to higher frequencies. It seems like the CCD& IOD voltages need to be tuned like a radio to find the right voltages to stop micro stutters. I got rid of mine when I did a BIOS update on my b650 aorus pro ax. Using F21 I was able to easily overclock to 8000-36-45 @ 1.5v, 1.05vsoc with XMP 6400mhz hynix a die ram. Now with 2200fclk i have zero issues& super smooth frame times using CCD @ 911mv & IOD @ 1047.
100% true, I've noticed the same behavior with my 7600!
@@fazendohistoria_ Have you fixed it with the right voltage settings? I found it way easier going 2:1 with 8000mhz. I could do 2200fclk easily (luckily on my 7800x3d with b650 aorus pro)
Exactly why I don’t run FCLK on my 7800X3D, haven’t had a single issue running Optimum Techs 7000 series settings, frame chasers a bozo 🤡
I dont want to OC anything on my 7800X3D. Do you recommend Curve Optimizer to negative 30 if that's the case?
@@puffyips of course he is a bozo. Guy is literally an intel fanboy
I had random lag until a bios update fixed the issue for me. The AGESA 1207 mentions this. It was intermittent and not too bad but so nice that it's gone!
Me with my 7800X3D just blazing through games 🤷
one of the top FPS gamers, Enders, uses AMD CPU. a 5900 series even, for gaming and content creation, he has really keen senses for any sort of dips, and says nothing is wrong with it. has won multiple tournaments for money on it. but he uses a 3090 for GPU. the way you should be looking for the dips, though is actually go play the game and look for those extreme dips while doing fast paced actions like whipping your mouse around. at 1440p or 1080p, frames uncapped
Even when he uses a 5900X he is still a lot slower with that then if he uses a 12900K instead of the 5900X on his 5900X. And true that fast mouse movement i a lot of times in cases with a lot of buildings and other places. Framechasers has a number of video's demonstrating it.
And even when the 5900X was not a bad CPU it was way worse then the 5800X that only has one CCD. So the AMDip is real. People just don't want to believe it because everyone else says so.
@@karlos1060”AMDip is real and is worse when you only have one CCD”
How uneducated can you be?
Also, learn how to proofread. You clearly know nothing about anything.
@@karlos1060 well that’s their fault for buying a 2ccd chip, 8 cores with 3d v cache is peak for now and a good while
"top FPS gamer.." are you on crack? Enders uses an aimbot. I might still have a link to the video in 2042 where he is caught 'snapping to target'. As a former BF admin, You guys worry me a lot when you cant spot a clear aimbotter...sheesh. It aint the cpu, hes just another kid with hacks.
Go in peace with this new found wisdom.
@@IANC4EVERyou might have a fundamental issue within your brain because i watched a couple of your (2 year old) videos and all it is is normal gameplay with bad interpolation of ticks in their spectator system
youve been in denial for 2 years
Had a 2700x and now a 5800x3D on the same socket and man owning AMD is way better than when I was stuck with my Haswell quad core waiting for Intel to deliver 😪
Are you compairing a 4 core 4th gen with a 8 core 2nd gen ryzen? Really? Take a I7-6900K, or some newer like the I7-8700K wich is in the same time or I9-9900K that came out 6 months after the 2700X you will see a huge difference in benefit for the Intel CPU. In games the 9900K destroys the 2700X. The 5800X3D is a good cpu. Not really fast in daily use like single core and multicore is not real fast but more then enough and it's made for gaming wich it does perfect.
@@karlos1060 9900K was Intel's last good CPU and it came out almost 6 years ago. Intel needs to wake up before AMD starts sandbagging and we have the same shitty situation I was talking about during Haswell
@@cosmic_gate476 That is absolute nonscence you are saying now. I am very happy with my I7-14700KF. It is very fast in single core and extremly fast in multicore.
And yes it can use more power. But in daily use it does not use more power then my 5900X i had. Even a little less!
And Intel is working on faster CPU's don't worry about it.
As i said my 14700KF is as fast as the fastest non Threadripper AMD there is on the market.
In cinebech R23 i get 2194 single thread and 35500 conistently in multi score. That is just very good! Only multicscore in Cinebench R23 uses a lot of power.
But in daily tasks and gaming it is very fine without using a lot.
@@cosmic_gate4769900K was terrible because it stole hyperthreading from the 9700K which is now way slower than the 9900K
@@karlos1060 you are complaining about the guys comparison being flawed and here you are trying to compare a 9900k a 500$ cpu with 2700x which is 330$ one, but now the difference between them is one is stuck to 9900k unless he changes his entire system while the other can just upgrade to 5800x3d which has gaming performance that matches 13th gen intel processors
When are people going to just enjoy playing games. 🙄
When good games come out that aren’t oldies
When a player can find out where to go, how to get there, and what to do when you get there.
Never, they keep the tab with all the GPU,CPU,RAM metrics open, and they jack off to them while playing
Why did you test at 4K and the highest graphical settings? Hopefully you realize that will fully load your gpu and not your CPU’s … the entire point of this video is to load the CPU’s … not sure why you would do that.
It’s a 4090. The GPU was at around 80% usage or less typically with FPS locks. Otherwise the results would have all been the same.
@@GraphicallyChallenged80% utilization at a fixed max fps for a 4090 still means you are not loading your cpu … to investigate an issue with CPU’s in games like this you need to run the games at resolutions and settings that are typically used by competitive gamers. The way you setup your tests does not load these CPU’s so how would you see an issue if it was there?
@GraphicallyChallenged bro your test is no good, he's right that is why you went silent.
if your having issues with stuttering disable resizable bar... it causes issues in some cases.
Should've reached out to Jufes to get his input. This is his creation.
Agreed. Atm none of them seem 100% believeble before they talk to each other.
@@TheBann90 I thought they were friends.
@@NBWDOUGHBOY idk... AMDip is more of a meme than just one person tho?
@@TheBann90 Definitely not a meme. Just has to be to the right condition to rear it's ugly head.
@@TheBann90It’s not a meme dude.. The sad thing is it’s real and people just think it’s Intel fanboy b.s.
4K for CPU test 😐🤦🏻♂️🤦🏻♂️🤦🏻♂️🤦🏻♂️
Usually you'd be right but we are looking at dips not average fps.
They are texting the cpu design integrity under “normal” use parameters for those chips….nobody who has an 7800x3D is playing in 1080p dude
@@cameronbutler6255 Actually 1440p man
@@christophermullins7163 you need to hit the CPU limit to see the dip guys, come on. 4k to find this issue is really working to disprove the problem.
I can’t believe people actually believe that lil weasel frame chasers it’s fkn wild
And I’m calling him that bc he knows what he’s doing he knows the hardware so he knows how to exploit it and make it look bad just to sell another product buddy is a sellout
@@puffyips why would falsely trying to discredit AMD matter to his business of building PCs and tuning consultations.
@@dinobot_maximize because he sells stuff for Intel
@@dinobot_maximize I personally have a hunch that Jufes intentionally sabotages AMD so he can sell his consults and his 6ghz 14900k apex encore bundles because the margins on those are far wider. A guy as knowledgeable as him can easily identify the underlying issues and fix them. I laughed when he played dumb and couldn't understand why his programs kept crashing to desktop and would stutter / freeze to a ludicrous level. He never showed his Bios setup which is why I have raised an eyebrow at him.
@@dinobot_maximize how dense are you lol? Intel? Ever heard of intel?
AMDip comes from Framechasers.
More like Framecrashes, people think he can make ultimate overclocks for their lcs, but even his pc is crashing live on stream habhaha.. And few people literally debunked amdip on cod and same settings with same location with fast movement, funnily enough is that they didnt have dips
DoofusJufes aka FramChasers Intel/Nvidia fanboy, fabricated this lie.
He sells "tuning" services for Intel, Nvidia systems, and also sells "tuned" Intel/Nvidia rigs himself.
He's the QuantumTV of techtubers.
was subbed to framechasers at one point but it kinda felt like qanon of tech channels so had to amdip outta there
Same here. He is a shill deluxe....
Tech Qanon is the perfect term.
The AMDip doesn't happen in all games, but it's definitely real. Battlefield 4 is one of the games I found that causes the AMDip. I went from 9900K to 7800X3D to 13700K because of the AMDip.
On latest win 11?
Yep it doesn't happen everywhere but it's very real
Did your 13700k already died from this insane voltages?
Frame chasers, an intel fanboy invented this nonsense and naaf pushes it.
Funny since naaf got rid of his all AMD system due to streaming problems his Intel/Nvidia system still has streaming problems but these are never attributed to his system but discord or something else lol!
@@davidg2731Framechaser's really small channel has convinced naaf and tgog that he is the greatest tuner of cpus that was ever born. It's ridiculous. The guy is a huge fanboy.
@@davidg2731Intel and nvidia is a great combination for content creators.
Hello! 5800X3D user here and I can confirm that I am able to use it with CO -30 on all cores totally stable in all scenarios (productivity+gaming). I might be somewhat lucky, but it is possible to undervolt X3D chips as well as a non-3D V-cache chip would. 👌 Silicon lottery is there for any chip, no matter the chip technologies used.
Yeah I have my 5800X3D undervolted with no issues at all but with way better power consumption.
-30 all core here as well, super stable. Tested for a few days straight never seen it crash, in gaming, browsing loads/low loads ( which a lot do not test for ) and productivity.
Why u undervolt tho ?
@@K1M01 It reduces the power the CPU uses which costs less in energy use and also makes the CPU run cooler.
@@K1M01 The CPU can run very hot out of box and will loose clock speed most of the time due to it. You can literally gain FPS as it keeps boosts way longer ( it can keep its boost indefinitely ) , its also way better for the chip allowing it to live a lot longer.
Perhaps someone at intel came up with "AMDip", It's kind of catchy! All they would have to do is say it a few times in the right places and gaming "experts" magically made it "exist"
7800x3d is best gaming gpu for 99% of gamers. That 1% are people who spent 5000$ on gaming rig to get every possible fps. I am still happy with 5800x3d..
Agree about overclocking the X3D CPUs, but I do not agree that undervolting is a problem. Obviously any time you mess with voltages you can reduce stability, but a well-tuned undervolt should be 100% stable and can get you up to a 10C temp reduction on the 5800X3D or 7800X3D. Very worthwhile.
Just need to thoroughly bench it, if you can do two 30 minute back to back cinebench stock, then 2 30m back to back on your uv and keep better scores without issues, it's most likely fine for longer sessions.
@@Haydenh127 UV's are usually less stable in lower intensity workloads and idle, not in stress testing. I would test undervolts by just using your PC normally for an hour or so. If it's stable, then you're good.
The caveat is when frames are locked. You should lock your frames anyway when you can since it drastically improves your 1% lows on either platform, AMD or Intel.
But for when frames aren’t capped & you’re pushing to their max, Intel typically has the advantage in 1% lows, albeit small.
The biggest advantages of AMD x3d > Intel this generation include easier ram timings, cheaper cost, “plug-n-play” best performance (if you don’t like to tweak), much better thermals/energy efficiency & the AM5 platform which should be supported a few more years.
Only way I can see Intel making sense is for high load productivity or enthusiasts who like to tweak their systems.
well at this point you should check may be if undervolting x3d's is an issue or rather mandatory😅
Take a look at Anthony Edwards' description of the differences between AMD and Intel design in his tech quickie RUclips channel
If there is some kind of frame dips on either company's CPU, it's due to a game specific bug, or you have some garbage running in the background, you can thank Microsoft for most of the frame dips usually.
Why didn't you use the 14900k?
Why would he? It's basically a oc'd.13900K.
@@curedanxiety1679 because he used current gen AMD .... if you want to compare to a 12900k then run a AMD 5950 or something... makes no sense
Everybody in the comments exposing themselves and their bias in the face of facts. Your settings being different may cause YOUR rig to be less stable; But the design of the chip itself isn’t inherently flawed, you’ve just seen it.
Both AMD and Intel have imperfect designs with their own strengths and weaknesses. I think it’s best to choose which CPU best suits your needs for the best price.
They are both great 👍
His benchmarks are with capped FPS and didn’t show any gameplay. The dip is not just 1% lows but the wild swings in FPS from 230-260 down to 150 it’s extremely jarring. And this is coming from someone who isn’t biased and have multiple AMD and Intel CPUs.
Undervolting X3D -30 is a godsend idgaf but I do keep expo off on my 7800x3d/ 32gb ddr5 6000 cl30 on xmp ofc
Hairchaser simply is incompetent and he can't for his life do proper testing and check if IF is stable (most of his dips if they really do exist come from IF clocked higher than it should be with error correction kicking in). If anyone experiences such things, you should fire up some memory stress test (i.e. ycruncher) and play music in parralel in your browser or with your fav player. You'd almost immediately hear if something is wrong.
Frame Chasers claims this dipping happens in COD, whipping the mouse around also may cause dipping. No body else seems to make a big deal about this though so I wonder how important it really is to most people!
No such results manifested during testing. These tests were done by 2 people actually playing the game on the same team side by side for a whole match.
@@GraphicallyChallenged I know which video he's referring to and someone pointed out it was because of his specific FCLK settings, they uploaded a video with proper settings and found no such issues.
@@GraphicallyChallenged are you just going to copy paste these two sentences over and over again? I guess you have simplistic, fanboyistic world view with A-vs-B, and these comments are on 'the wrong side' so you feel a copypaste is all they deserve. Real classy. Very weird is that the point of the person who you are replying to doesn't even care about dip this or that, it's about if people even REALL care (most people don't know what a Framechaser is, or what a GraphicallyChallended is). They don't - mostly no one cares about this "narrative I see in my comments". Well, the AMD fanboys will be happy, and when you make any kind of fanboy happy with some "I didn't really need to make this" kind of video (it's not some new product, or actually-damaging narrative), that doesn't really speak well of you.
Next, how about a video that, actually, no one cares about stupid "dead platform" arguments (including AMD fanbaby's who will ignore it when someone ask about getting AM4 right now). Many narratives are stupid, or might have some minor point, but not really anything most people will ever care about or should care about.
How about reminding people that AMD deosnt win all gaming bench's, and doesnt win all productivity tests - they win on, to some degree price, but mostly watts. Which, actually, most people dont care about, they just pretend to online so virtue signal or win internet points in fanboy wars.
Also, for fun - I guess a lot of benchings from GN are lies.
Yes but it was near a certain building structure / scene. But yes it does happen and I’ve experienced it on my X3D CPUs …. 5800X3D and 7950X3D (running in both modes as 7800X3D and 7950X3D). Ive also posted in these comments about my experience.
FrameChasers is not a reliable source of data or benchmarks. His findings do not match those of known-reliable sources.
The idea of AMDip is not really something you can catch with benchmarking. AMDip is all about fast movements with the mouse in random areas of games. It is not just tied to the CPU either it is common in the gpus as well. If you have both in your system with fast twitching movements it can be more obvious, but can be also be found in just the AMD cpu’s.
What you said about overclocking AMD cpus i think really is the main point. With Intel it seems that overclocking doesn’t affect its stability it makes it stronger, when done correctly. And with AMD it gets more unstable no matter what. so that could make the dip even more obvious any why people are talking about it so much these days, they are trying to OC their AMD cpu’s and ram.
The intel cpus just rapidly degrade and crash games after few months of use instead lol.
Oh man, someone’s about to whip out their note pad and make crazy faces while annunciating loudly.
Don't forget "right" after every sentence.
I can’t wait lol 😂
I have had 4x Ryzen CPU's and the DIPS are not a issue, Its the Audio, Crackling and poping unfixable got this on 3/4 ryzen systems here. Something to do with ryzen infinity fabric voltages, But I cant fix my sets. Will be buying intel in the future. Never mind the memory issues on Ryzens 2nd gen.
Thanks for looking into this.
Ya, I am happy he looked into this rumour and set us straight.
lol he didn't look into it. He said at the start of the vid it didn't show up in standard benchmarks and then goes and shoes standard benchmarks. The dip happens when you move the mouse too fast in the game not in the automated benchmarks.
@@Roman00744 He made a video about it. I consider that looking into it.
@@Roman00744 So you're saying that if there's a dip, an actual stopping of movement, that that wouldn't have an effect on the 1% and 0.1% lows?
@@pf100andahalf No I'm saying it's not happening during auto benchmarks, you need to actually play the game and do fast movement with the mouse for the dip to appear.
Last system I built used a i7-7700k and Asus was a sure buy. Things seem much more confusing now.
You really should not test CPU consistency at highest settings in 4K. Also you should get the most complex difficult scenarios in a gaming title, that's where AMD goes down. If there are 10 people around you shooting, throwing grenades,... Then you'll see a big drop in 1% low with AMD when a 14700K/14900K stays stable :)
I noticed the delay opening programs in windows frame chasers was on about with my 5800x3d (I was using windows 11). I recently reinstalled windows and my system is so much snappier so wondering if something has been updated that helps. In gaming I’ve never really had an issue. 1% lows were significantly better with my 5800x3d cpu vs my 5800x.
What resolution do you play and what GPU?
@@charlie7mason 3440x1440p, at the time I swapped from the 5800x (had a use in another build for it) I was running a 3090. I’m currently running a 4090, with it.
@@curtismariani6303 Thanks, I'm really tossing and turning about upgrading from the 5800X to the X3D on a 7900XTX playing at 4K. It's a tough one. Benchmarks don't show much if any difference at that resolution.
@@charlie7mason I’m not sure it would be worth it. It improved my 1% lows on warzone by somewhere around 10-15% the 5700x3d is coming out for around $250, if you could sell your 5800 for $150ish then maybe worth it for $100. But I wouldn’t spend more that that. I’m holding on to my processor for another year at least, I think. Unless I go for a 5090, I don’t plan to jump to the new socket anytime soon.
@@curtismariani6303 Yeah, I figured that as well. Since I'll be gaming at 4K for the most part, I don't see the need to jump on it any time soon. Plus, I'm not switching sockets for quite a while yet either, so I can wait for deeper discounts in the future if I really feel like getting it.
And what about AMD cpus with 2 CCD and without vCache?
I’ve had a very mixed experience with these depending on the motherboard, BIOS, and RAM.
Sometimes it’s great, sometimes I get lots of stuttering depending on the hardware I pair with them.
Frametime spikes on multi-CCD, non-X3D CPUs *might* actually occur, because of the interconnect latency. I haven't seen any evidence or widespread reporting of the fabled "AMDip", but I do see ever so slightly worse 0.1% Lows on the 7900X and 7900X3D compared to a single-CCD design like the 7800X3D or an 8-core CCD design like the 7950X3D.
@@GraphicallyChallenged then why show half truths
X3d still dips. Not as bad, but it does dip.
@@mikemoore5270 No, it doesn't. No one with any clout has shown the fabled "AMDip". The only people pushing that narrative are FrameChasers and his cronies, who are NOT a trustworthy source of information. He's a fraud. Plain and simple.
intel fanboys are not gonna like this video haha
the dip is real, but its not a big deal. its only under certain circumstances in particular areas of games. if you went to these particular areas in games and hopped on one foot while touching your nose with your right index finger, you could make it studder. im on a 9900k right now, and if i had to upgrade i would probably go with a 7800x3d.
This dip is not real, user error is, early adopters are idiots and people who just say what they hear are spreading misinformation 🤦♀️
Great work checking not just 1% lows but several other "min" criteria along the way. Steve at HUB also did a superficial investigation maybe a month or two ago into any potential empirical substance behind the narrative and also found nothing.
FrameChasers is primarily an Intel/Nvidia fanboy appeasement channel and he's been promoting this "AMDip" narrative the most as far as I've seen. I prefer to avoid rewarding his content with views or any other sort of engagement.
A big part of his business model involves selling very expensive tuned Intel motherboard, cpu, ram combos. Using an AMD 7800X3D combo does not require any complicated tuning so making his pricy services unnecessary!
Never heard of or seen it happen with my three-month-old 7800x3d with OLOy 32GB DDR5 6000Mhz. No OC applied to the CPU (I think the max that can be gained is 5050 Mhz no matter what), only to my MSI Ventus RTX 4070 Ti, which is mildly set by MSI AB. No complaints, no dips, no errs...
8 cores don’t dip. It’s the 12 and 16’s
Not even -30 all cores, 85w cap and 85° temp cap. These three things is peak for 7000 series
dual ccd dip because games might switch between ccds@@mikemoore5270
Seriously I have AMD Ryzen .. online gaming since like always ... not a single issue! FPS runs at cap, no drops, no spikes, no overloads or overheat etc ...
Benchmarking cpus at 4k 💀💀💀
AMDIP is from frame chasers YT channel. He is comparing max overclocked 13900K at 6Ghz with 8Ghz ram vs 7800X3D also tuned at max. Intel can gain alot of performance from OC and tuning. AMD on other hand not really.
Btw testing cpus at 4k not great idea.
You forgot to mention frame chasers purposefully uses FCLK on the X3D which can cause issues aka self inflicted dip he knows what he’s doing
Nothing new. Been experiencing strange performance drops since the Athlon x2 days. My current systems are AM4 3950x and AM5 7900x3d.
This isn't how to test this.. you need to be running realtime frametime + mouse whip = you can visually see the spikes in realtime frametime graph.
This is a CPU test, not a GPU test. That wouldn't change the outcome.
Yes, you need to physically use it to see the constant jarring 💯
Potential issue with dual CCD, non 3D-stacked AMD CPU's. Tests with a Single CCD and 3D Cache CPU. Congratulations. You have proven nothing
There's some truth AMD were a bit inadequate for high FPS games for a stretch. A friend was having trouble with Apex Legends - no idea why that game's so CPU demanding - anyhow the frame rate was dipping. He said people with much better systems than his also dipped in Apex. I got him to replace their Ryzen 2600 with a Ryzen 3600. No more dips. He could drop in a 5600 now or even 5800x3d, but he's happy with what he's got. The more difficult problem is if AMD keep making improvements what will we call the Intel 'dip'? Intelimping? Inteloosing? I can't find anything that flows.
The people claiming "AMDip are the people who don't understand AMD's fundamental architecture and either just like to pump numbers and voltages as high as they can like its an intel chip or have an incorrect bios setup. AMD's architecture doesn't allow for that kind of tuning while remaining stable. Basically the important thing to note between AMD and INTEL is that for AMD, tune for better synchronization and less latency (while remaining fully stable) rather than the highest speeds and voltages you can get on memory, core, etc. A stable 6000mhz kit with tight timings is better than a 6400 kit with more latency. From my experience on ryzen 3000-5000, Infinity Fabric should run 1:1 with Memory. so for example, if you're running say a 5900x with ddr4 3600 cl18 kit, Fclk should be at 1800mhz to be 1:1 with Mclk. On my 7800x3d, I have found the platform to be smoothest when IF runs 2:3 with memory which lines up with tech reviewer's recommendations perfectly. This is why people and channels have repeatedly recommended a 6000 (preferably cl30) kit because any 7800x3d can run 2000mhz Fclk stock without any manual tuning. My personal recommendation for 7800x3d is to buy a 6000cl30 kit, enable expo (assuming the manufacture's expo profiles themselves are stable at this point), offset -10 to -20 all core undervolt in PBO, and leave it there. Keep in mind that an all core offset can be stable on some cores but not others (This is why a -30 might technically work but actually be less stable). Verify that FCLK is at 2000mhz, Re-Bar on, and call it a day.
Meh, probably just some Intel fan boys trying to discredit AMD as its getting more & more popular because of the bang for buck u get from AMD vs Intel.
Amd haters are something else lol 😂
You didn't compare with the 7950X3D which has 2 CCDs vs one in the 7800X3D
wouldnt make a differnce in games if you use it correctly, ie using process lasso to set games to the 3d cache ccd
You are comparing it with 2 generations old 12900k, on GPU intensive 4k resolution and locking FPS. Terrible test
I appreciate the video, but I'd like to see this test redone at 1080p Max settings to ensure the 4090 is never the bottleneck, I get you capped the fps to prevent the gpu from maxing out, but at the same time that would alter the consistency of the cpus frametime performance, it would fluctuate a lot more if the cpu was always the limiting factor, plus would have been a more even comparison to try and compare it to a 13900k as I believe that would would have been the CPU to compare the 7800x3D to when it was released.
Little of column A, little of column B.
As a hardware collector, and hardware admin that manages ~1500 devices at work.
Yes. AMDip exists, but iDip also exists
IMO Intel generally takes less work to alleviate the dips, short of the X3D parts.
Also AMD screw you for cutting the cache in half on APUs Know what will save battery life on laptops? Not having to go out to RAM as often. Know what causes a CPU to go to RAM more often? Cutting the cache in half
It ticks me off that i can get the 5700G to 2500 FCLK/MCLK/UCLK(this means DDR4 5000C18 in gear1), but its still not quite to the performance of the 5800X at 3600 because of that lack of cache. Well sometimes its better than the 5800X3D, but not that many use cases prefer higher bandwidth over more cache; mostly older games, but not so old that the whole game can fit into cache. But also if a game is so new that even X3D isnt enough cache, the 5700G can claw its way back.
Weird little tidbit, i had to turn on adblock to be able to see comments, including the one i posted, i even tried closing and re-opening the browser.
Also, RUclips, if i'm typing a comment, and you decide to play an ad, dont steal my cursor and make me click on the ad when i hit the space bar, i ended up having to type the above comment twice.
Speaking of FCLK/MCLK/UCLK, i cannot wait to get my hands on an 8700G and a good topology iTX B650 board(maybe X670, but B550 was better on AM4 for memory)
If i can get 2500 on the 5700G and 4750G with 7nm, i am excited to try and hit 3200-4000 on the 4nm 8700G for 6400-8000 in gear1 Hopefully FCLK and UCLK can keep up, not quite as important on AM5 supposedly(i dont believe it) but still a noticeable difference for the iGPU if FCLK/UCLK are 1:1 with MCLK
For instance, i have a Z1X and a 7940HS.
Despite both of them supposedly pulling 29w from the socket, the 7940HS is often faster with 5200 in 1:1:1:1 than the Z1X with 7500 in 1:1:4:4(FLCK/MCLK at 935Mhz but MCLK apparently reporting at 2750 somehow, even though 1:1 would be 3750, and 2:1 would be 1875 i honestly do not know how 2750 works with 7500MT/s, must be dropping to 5500MT/s and the software just isnt catching it)
This may also be due to the 7940HS having 32GB of RAM with an 8GB 'partition' for the GPU where as the Z1X only has 16GB with 3GB for the GPU. Some games this doesnt matter, and would run fine with 128MB for the GPU, other games with better optimisations will try to stay within this virtual partition dumbing down settings and hitting the SSD harder by purging and re-loading assets more frequently.
More cache means higher temps...which affects laptops much more than desktops.
@@GSP-76 more cache, means less RAM usage, means lower picojoules per bit
You know cache consists of transistors, right?
You also know that for a given process node and a given die size the number of transistors is limited, right?
So, if a CPU with a bigger iGPU has to make do with the same amount of transistors, how is that going to work? It's easier to convert those cache transistors into iGPU transistors than cut something else.
How could an APU have the same core count, thread count, transistor count, amount of cache and clock speeds with a bigger iGPU without a node shrink? And how could it have all the specs of a normal CPU with a bigger iGPU and cost the same or less?
It doesn't make sense.
Bro used 12th gen intel on ddr5 6400…. See guys you can’t trust other on the net.
Proud owner of 7800X3d
Same brother and no issues either it’s been fkn great performance with lower temps than the 5800x3d
Why is it that video existing because "waaa, I see comments" are always at least a little sus, if not entirely dogshit? Trust me brah, here are some barcharts!
Also - why the 12900K and not the 14900K?
i have a 7800x3d + 4090 + DDR5 6000 cl30 with those old buildzoid safe hynix timing but i also have Asus's CPU overclocking profile turned on forgot what it's called but brings the cores to 5.2 lol might have to revisit that overclock profile.
I haven't noticed an issue since the 7800x3d released so now i'ma start testing games and looking to see if there are massive frame drops with fast movements.
if this only happens on competitive games like overwatch or csgo etc where people play at 1080p then yeah i'm probably not going to see a problem since im a graphics whore and play at max settings 4k lol maybe if i'm playing a game that isn't gpu intensive at those settings but i have to find one like that since im not testing the games at 1080p since i'm never doing that.
So no all core negative CO and maxed out Loadline??
Could be using a dodgy powerpack or one that has capacitor issues that does cause performance dip issues noticed that with a few retro Athlon builds while the chips was perfect the PSU's were aged after replacement there was zero issues with the builds also bloat ware or software like some anti virus scanning soft wares that can interrupt other software running it will drag a PC down
Does AMD use fewer checksums than Intel and if so, does it result in more execution errors?
Please define "execution errors" (although I suspect that you can't). Also, checksums are a way to detect and potentially correct memory corruption, but they don't stop any memory corruption from happening in the first place.
audio? its not practical but imagine your favorite part of song/audio change places/tune/quantity itll become unreliable to score or recommend since idk what causes those
not practical in sense most would find it no biggie, same as recording reliability, and true i can just score low
its just how stage monitors work, produce sound as accurate, not as good/pleasing, pros need those
Seen people mention his channel a lot recently. They forget to mention that he sells systems that mainly consist of intel. Likely just cherry picking results to sell more. Gamers nexus could resolve this
If you really think that, you don’t know what you’re taking about.
@@LegionGamingTV i mean this video is showing better result with amd, and the one that he mentions amdip is just warzone. Again, gamers nexus could have a good comparison if he doesnt already.
actually he does have it, look at the 14900k review. 7800x3d beats the 14900k in average, and is usually better or the same for 1% low
@@SeventhCircle77 Gamers Nexus doesn’t test Warzone for one, for two he doesn’t show any gameplay of what the AMDip actually is. Do you really think gamers nexus sits there for hours playing games? They have a set benchmark just like everyone else and if they don’t find any deviations in the 60 second to 3 minute run, you won’t hear about it. It doesn’t mean lower 1% lows, the AMD CPUs bounce from 250 fps to the 1% all the time instead of it being a gradual decline. It’s extremely jarring and feels like a lag spike or stutter when it happens. Going from 250fps to 140 consistently while using a 240hz monitor makes for gameplay that isn’t ideal for playing in tournaments and for big $100,000 games. If you’re an average gamer that doesn’t understand what’s happening you won’t know, and what’s worse it doesn’t show up on a bar graph. So sorry bud, you don’t know what you’re taking about.
@@LegionGamingTV sounds like a skill issue tbh my frame rate is dead steady. There's plenty of frame time graphs that show it does fine as well. My 1% low might be 40 fps below my average in most cases
@@SeventhCircle77 lol k
you need to look at live frame time graphs while gaming to ascertain real stutter these static graphs and 1% and .1% lows are essentially meaningless.
I could perceive 0 appreciable differences besides AMD having higher fps. Between that and the data I think we can safely say this case is closed.
Stutters would show up in the 0.1% lows if they were at all prevalent.
So called over clockers hitting power limits and not understanding what is really happening...also jedec memory timings are needed for fair comparison
All I know is AMD better be prepared to stop using TSMC if rumors are true about China preparing to take Taiwan by force soon while the US is distracted with middle east stuff and the open border.. I have has a 7950X3d and see no issues gaming so I think this is some kind of Intel ruse to get people to switch to Intel.
TSMC isn't going anywhere...
@@GSP-76 Umm, huge security risk using TSMC still with China in control of it, doesn't matter if it's not going anywhere..
@@bes12000 no it's not...and as I said earlier TSMC isn't at risk or falling into China's hands. It's not just AMD or a few popular companies that use their fabs. The entire planet does for thousands of different chip fabrication needs ranging from pennies to hundreds of thousands. No one is going to harm TSMC in any capacity and China themselves know this very well. In another 5-8 years competition will likely have caught up anyway and they're going to pivot to graphene. Silicon is close to EOL.
Also, AMD isn't the only one using TSMC. Intel is even using them because their own fans are so far behind. Nvidia also uses them along with Qualcomm...and that's just the top of the iceberg.
@@GSP-76 Not paying attention are you, fine, we will see what happens, if it happens don't say I didn't warn people, if it doesn't fine I was wrong, there are a lot of signs pointing to it as being a high possibility right now, China's not stupid and the US is weak atm with a weak leader who caters to Iran, and China, China's not our friend they have infiltrated the US with the open border, they know how to cripple the US, none of this is rocket science and not exactly new news either, notice all the AMD CPU's saying "Made in China", they are made in Taiwan, lots of products made in Taiwan have made in China on them now when they used to say Taiwan.
You should worry about something else cause that ain't happening. Taiwan is extremely important not only to the US, but to the rest of the world as well. TSMC is the world's biggest semiconductor chip manufacturer and it's a vital supplier to the US and other Western nations. It produces more than 90 percent of the world's highest-tech chips. The US simply cannot let a corrupt nation like China get ahold of Taiwan. The US will do anything to prevent that from happening even if it means going to war. TSMC isn't the only important large company in Taiwan. There is also ASUS, Foxconn, Pegatron, AUO, MediaTek, Quanta, Novatek, Acer, MSI, Gigabyte, ASRock, Realtek, InnoLux, BenQ, AOC, ADATA, etc. All those companies play a major role in the global economy. Trust me, nobody is touching Taiwan.
In another comparison for Intel Haswell Family, a i5 4690 presenting 1% low & 0,1% low score points way better than his sibling i7 4790
I've seen this "AMDip" thing being used a lot on framechasers (pretty much every review he does on AMD CPUs). The comment section frequently points to the issues you have mentioned in the video with instability, yet he never seems to respond or do any benchmarks without oc. I'm not accusing him of lying, there might be issues somewhere at some point in specific games in particular scenarios, it is the nature of how complex these architectures and the software we run are (Digital Foundry has discussed and shown this). Both intel and AMD have their advantages but I'd love to see him or anybody else do his benchmark run on a 7800X3D without the oc to test if it really is caused by that. My guess is that the spots where AMD has worse 0.1% lows are caused by games/workloads that scale better with clockspeed/ipc or memory speeds (since intel can utilize oc memory better) or AMD's c-states/general power-saving features that run more aggressively than intel's I believe.
I'll accuse him of lying. He clearly says whatever it takes to sell his products and services and blow smoke up his "supporters" ass. He is a slave to the income he gets from RUclips and shit. Fkin clown imo
Framechasers is not a reliable source, benchmark channel, or authority on... well, anything. He has a vested interest in Intel CPUs winning, since he sells a custom tuning service for Intel CPUs and platforms.
His numbers and findings do not match those of any known-reliable source, such as HardwareUnboxed or GamersNexus. He should be treated the same as we treat userBenchmark: with skepticism and ridicule.
@@marxman6896 been saying this. He says what sells his products and services. He is not the genuine, for-the-ppl source that he claims to be.. far from it. His "supporters" send him money so he will keep saying what they want to hear lmao its pretty easy to see as well not sure how people fall for it.
@@christophermullins7163 yeah it was particularly hilarious to me when in one of his videos he described himself as “too principled” or something…lol, just like every time a certain celebrity-turned “politician” says “believe me”. Oh, the used car salesman antics. So believable.
The dip their talking about happens With cpus that have the two chip infinity design 5800 &up (7×××also) because of the communication between the two chip infinity fabric design in which the problem is really in the non 3D versions. Which is why he used a 3d version with a generation behind intel. Not sure. What the problem is with this guy being deceitful towards the public is. I've seen this issue replicated on multiple test and not just in online gaming does this happen. Gamers Nexus is one site, it's the easiest testing results to find on this topic.
@GraphicallyChallenged As far as I know, this "AMDip" came from Frame Chasers, a guy that is bad at RAM OC and tells people that their RAM OC is fine as long as it is stable in the games they play (dedicated stability tests do not matter...). Badly-tuned RAM can lead to stutters and inconsistencies in games and other applications.
I disagree with telling people not to overclock/undervolt their X3D CPUs though, you just have to do it right, and no, they do not really operate at the edge, they often have very high Curve Optimizer margins.
The 7700 is much better as far as non 3D cpus because of the single chip infinity Fabric design
CAN I do ram memory xmp enable atleast ? 5:30
6:30 DOCP is ASUS's name for Intel XMP...AMD EXPO is named as EXPO.... I have an ASUS X670E Hero with a 7800X3D and i can tell you that XMP or EXPO or manufacturer or QVL list doesn't matter whatsoever. Buy which is the cheapest.
Oh, and come on, no one is playing competitive games at 4k.
You should ask Jufes about AMD Dip
Just because you didn't find it doesn't mean that it doesn't exist. PC issues can be cause by one setting being changed.
So your admitting it can be your settings and NOT an inherent design flaw with the chip itself🤦🏽♂️
lmao..yeah the only person i see shit on AMD is that frame chasers dude..i dont get the logic behind his followers..often they cling to..i wanna run whatever ram speed i want! like okay @@cameronbutler6255
@@cameronbutler6255 If any setting causes it then it's an inherent design flaw in either the cpu or motherboard.
@@JamesSmith-sw3nk sure just like e-cores causing atrocious mins in some games. Inherent design flaw by Intel. Must be.
@@JamesSmith-sw3nk Don't forget Windows too... they have done questionable choices for a while.
For me, the biggest problem with Ryzen is its idle power usage.
When you doing normal things like using the Internet, watching YT, or doing anything outside heavy CPU loads Intel CPUs are using 8-9W of power while AMD is using 40-50W.
The only huge difference is when you are doing a stress test with 100% CPU utilization but this never happens in the real-life scenario and with normal PC usage.
Also worth mentioning is that this amdip happens on 2 ccd CPUs where processes are wrongly distributed between them.
Stop the cap
Reading these comments my Ryzen 5 7500F is using 25W.
literally using 1.3 W right now with 2 monitors typing this comment on a 7800x3d lmao
I really don’t mind my 7800X3D pulling 29/30w idle when it hardly goes over 65w gaming even with a cap only at 85w, it will hit that 85w in cinebench ofc. I just love seeing it hit 5ghz gaming knowing it has that 3Dvcache
Running it against an older generation is a bit of a bias in and of itself by the way. As the range will always be bigger if it scrapes the ground. So you kinda showed which team you are on so to say. Making this test not believable. Perhaps also then indicating that you handpicked the games and/or settings.
nah, both side indicating hand pick games and setting anyway
Yeah but he’s probably not going to go buying a 14900 just for four benchmarks, is he? The architectural difference between 12th, 13th, and 14th gens is almost nonexistent anyway, and the allegation is this is something Intel magically never exhibits and AMD magically does. I’d say the evidence refutes the nonsense.
Its has to be viewed during gameplay cod warzone usually show these results of the dip on the newer generation and the older one as well
No such results manifested during testing. These tests were done by 2 people actually playing the game on the same team side by side for a whole match.
Not trying to attack nvidia fans but a nvidia fanboy probably made this up 😂
This topic is a fanboy topic. Move on. Nothing to see here.
Look in just this short run, it shows that all in all AMD remains typical when acceptable parameters are used it's just still typical there is a better chip there's all that don't win and truly a waste of time to examine anymore, unless you just like self torture.
who really cares both are good especially for gaming you dont need the fastest most wont use a 7900xtx or 4090 take whats in your budget, these comapnies poop on you guys
7800x3d is worst cpu to test. 2 ccd cpu needed. Especially 5900x.
That sounds like some Intel FUD bullshit
Amd am5 is supporting multiple generations. Am4 was supported for 7 or 8 years. Upgrading is just dropping in a new cpu after a bios update and maybe a new cooler if you go higher than an r5.
Intell upgrade costs you a whole new platform except ram and that costs you double.
Ofcourse budget oriented gamers are going amd. Most of us ain't earning 100k a year. Most gamers are blue collar workers and we do not want to spend extra money if we can spend it on our kids.
Intell shills and paid commenters ofcourse have to badmouth amd. Its at least 50% more expensive and you have to buy a new mobo. This is a lot of money intell is missing out on.
DOCP / EXPO
yes
I'm sorry, but what does the chart mean by "Min"? How can the actual "Minimum" be higher than a 1% or 0.1% of the chart?
It’s the standard terminology used … min doesn’t mean minimum value in dataset.
@@blackbirdpctech well.... What does it mean here???
Curious why you would use a 12th gen intel?
Also, when you say dont oc the 7800x3d, are you talking about not using PBO?
12th gen is the most reliable. 13th/14th gen have voltage issues.
@@julmdamaslefttoe3559 they don't have voltage issues. Motherboards were just pushing too much power. Intel spec is more than stable
@@fallenlegion1828 failure rate and people complaining is slightly higher. Id get a 14th gen Chip if i wasnt worried of the crashing issues.
@@fallenlegion1828 im looking at a 12400f for £98 or a 12600 for less than £150.
Intel fanboys are jealous of AMD.
because Intel generations 13 and 14 are Alder Lake Refresh with extra E core.
and some news is that Intel will launch another Refresh for desktop PC CPU in 2025 ~
🤣🤣
QVL lists dont matter
Not really but Expo Ram vs XMP ram 100% does
Why are you using a 12900k instead of a 139/14900k, both of which have more cache and slightly higher clock speeds. Also the AMDip occurs on cpu's using multiple CCD's.