- Видео 57
- Просмотров 237 434
Hardware Museum
Чехия
Добавлен 29 окт 2018
Welcome to the Hardware Museum RUclips channel.
I have quite a big collection of PC hardware. Mostly Video cards and processors, but motherboards and other HW too.
My specialization is benchmarking and the most of all gaming performance benchmarking. Of course there are many reviews of pretty much any video card or processor online, so why bother to do that again, you might ask. Well, the answer is simple - to focus on aspects you can't find in the old reviews. Here you can find benchmarks of games that came years after the HW in question. And to get really in-depth look on the tested HW, I am measuring the following:
Frametimes
GPU Load
GPU Temperature
Video RAM Load
CPU Load
CPU Temperature
CPU Power
System RAM Load
Power Consumption
Noise
Video capture of in-game footage
All this is packed into the GPU Duel, a short video in which you can find out how two selected video cards perform against each other. At some point in the future, there will be CPU Duels too.
I have quite a big collection of PC hardware. Mostly Video cards and processors, but motherboards and other HW too.
My specialization is benchmarking and the most of all gaming performance benchmarking. Of course there are many reviews of pretty much any video card or processor online, so why bother to do that again, you might ask. Well, the answer is simple - to focus on aspects you can't find in the old reviews. Here you can find benchmarks of games that came years after the HW in question. And to get really in-depth look on the tested HW, I am measuring the following:
Frametimes
GPU Load
GPU Temperature
Video RAM Load
CPU Load
CPU Temperature
CPU Power
System RAM Load
Power Consumption
Noise
Video capture of in-game footage
All this is packed into the GPU Duel, a short video in which you can find out how two selected video cards perform against each other. At some point in the future, there will be CPU Duels too.
GeForce GTX 560 Ti OC vs GeForce GTX 480 - GPU Duel (ep. 38)
In 2010 when Fermi based cards launched, the GTX 480 wan't considered to be very successful. Later when GF104 (GTX 460) appeared, there were speculations if using fully active GF104 wouldn't have been better for gaming purposes. So I took the opportunity to see how this turns out.
I used the best case scenario for the smaller Fermi - GTX 560 Ti (GF110) overclocked to 1 GHz and put it against stock GTX 480.
Video Cards tested:
GeForce GTX 560 Ti OC (Gigabyte SOC) hw-museum.cz/vga/346/gigabyte-geforce-gtx-560-ti-oc
GeForce GTX 480 (Reference) hw-museum.cz/vga/301/asus-geforce-gtx-480
Content:
0:00 Introduction
0:28 BioShock Infinite
2:51 Crysis 3
5:58 Dirt Rally
7:43 GTA V
9:37 Metro: Last Light Red...
I used the best case scenario for the smaller Fermi - GTX 560 Ti (GF110) overclocked to 1 GHz and put it against stock GTX 480.
Video Cards tested:
GeForce GTX 560 Ti OC (Gigabyte SOC) hw-museum.cz/vga/346/gigabyte-geforce-gtx-560-ti-oc
GeForce GTX 480 (Reference) hw-museum.cz/vga/301/asus-geforce-gtx-480
Content:
0:00 Introduction
0:28 BioShock Infinite
2:51 Crysis 3
5:58 Dirt Rally
7:43 GTA V
9:37 Metro: Last Light Red...
Просмотров: 202
Видео
Core 2 Extreme QX6850 vs Phenom X4 9650 OC - CPU Duel (ep. 10)
Просмотров 4594 месяца назад
The first Core 2 Quad was introduced in late 2006 and it crushed all competition at the time. AMD had their quad core design in development but that was very far from ready in 2006. Eventually the first Phenom appeared a full year later, in November 2007. These CPUs had a lot of issues - like the famous TLB bug and low clocks speeds. These problems were mostly solved with late B3 stepping, show...
Athlon X2 7550 vs Core 2 Duo E7200 - CPU Duel (ep. 9)
Просмотров 5219 месяцев назад
This time let's compare two mainstream solutions of year 2008. At the time AMD finally managed to deliver fixed B3 stepping of the original Phenom. The whole lineup was based on the same quad-core design with various amount of cores (de)activated. So Athlon X2 7550 is one of the slower models with only half of the cores active, but still full 2 MB of L3 cache enabled and frequency of 2.5 GHz wa...
Core 2 Extreme QX9770 vs Core i7-975 Extreme - CPU Duel (ep. 8)
Просмотров 2,9 тыс.Год назад
This time I took the fastest CPU from the Core 2 family - the Core 2 Extreme QX 9770. It runs at 3.2 GHz and is the only Core 2 using FSB at 1600 MHz. On the other side I picked the fastest model of the first Core i7 generation - the Core i7-975 Extreme. Bac in 2008/2009 when Nehalem was new, the reviews didn't show much improvement in gaming performance, so let't see how it is today. All the t...
Radeon R9 290X OC vs GeForce GTX 780 Ti OC - GPU Duel (ep. 37)
Просмотров 587Год назад
Radeon R9 290X and GeForce GTX 780 Ti, the flagships of 2013. Both cards overclocked to the max. This time I tested more games, older ones and rather modern too. Let's see how the famous "fine-wine" turns out. :) Video Cards tested: Radeon R9 290X (Reference) hw-museum.cz/vga/347/sapphire-radeon-r9-290x GeForce GTX 780 Ti (Gigabyte WF3X) hw-museum.cz/vga/327/gigabyte-geforce-gtx-780-ti Content:...
Athlon 5350 vs Core 2 Quad Q6600 - CPU Duel (ep. 7)
Просмотров 676Год назад
In 2014, AMD came up with interesting platform - Socket AM1. It is pretty much lowend mobile SOC modified to PGA socket and fit desktop boards. As you can imagine, the power consumption of such solution is great - even with GeForce GTX 1080 in my test system. The question is - how it performs compared to good old Core 2 Quad Q6600? All the test scenes are 100% CPU limited. Testing resolution is...
Core 2 Extreme X6800 vs Athlon 64 X2 6400+ - CPU Duel (ep. 6)
Просмотров 870Год назад
After many years of struggling with the NetBurst architecture, in mid 2006, Conroe was finally ready for market. I took the fastest model of the initial wave - the Core 2 Extreme X6800 (running at 2.93 GHz). This one was by the way the last dual-core "extreme" CPU. As the competition I selected the Athlon 64 X2 6400 . It was released a full year later, but still based on the old K8 architecture...
Phenom II X4 940 BE vs Core i7-920 - CPU Duel (ep. 5)
Просмотров 1,6 тыс.Год назад
Let's get back to early 2009. At this time AMD was ready lo launch the long awaited Phenom II. At first there were two SKUs ready - X4 920 and X4 940 BE, both for socket AM2 . After the not so good first generation Phenoms it was definitely step forward. I used the fastest model - X4 940 BE. As the competition I took the Core i7-920 which was also very new at the time (launched in November 2008...
Phenom II X6 1100T OC vs Core i5-750 OC - CPU Duel (ep. 4)
Просмотров 3,1 тыс.Год назад
This time we will see the battle of the AMD's fastest K10 consumer CPU - the Phenom II X6 1100T and the mainstream Nehalem part - Core i5-750. Both are overclocked to the stable maximum, which is 4.1 GHz for the Phenom and 4.0 GHz for the i5. The CPU-NB of both CPUs is also overclocked - 2.6 GHz for the Phenom and 3.2 GHz for the i5. All the test scenes are 100% CPU limited. Testing resolution ...
Athlon X2 7850 OC vs Athlon 64 X2 6400+ - CPU Duel (ep. 3)
Просмотров 1 тыс.Год назад
The main goal of this episode is to compare AMD K8 and K10 architectures at the same frequency (3.2 GHz). Here the K10 is represented by the Athlon X2 7850 BE - the first generation of K10. Athln 64 X2 6400 is the well known Winsdor core. All the test scenes are 100% CPU limited. Testing resolution is always 1920x1080 with 16xAF and vsync off. All the videos are recorded using dedicated PC with...
Core 2 Duo E8400 vs Core 2 Duo E6850 - CPU Duel (ep. 2)
Просмотров 1,9 тыс.Год назад
This is detailed comparison of CPUs gaming performance. Very much like the GPU Duels. All the test scenes are 100% CPU limited. Testing resolution is always 1920x1080 with 16xAF and vsync off. All the videos are recorded using dedicated PC with capture card, there is zero performance impact on the tested CPUs. CPUs tested: Core 2 Duo E8400 [2C/2T, 3.0 GHz] hw-museum.cz/cpu/459/intel-core-2-duo-...
Core i3-530 vs Phenom II X3 710 - CPU Duel (ep. 1)
Просмотров 1,1 тыс.Год назад
This is detailed comparison of CPUs gaming performance. Very much like the GPU Duels. All the test scenes are 100% CPU limited. Testing resolution is always 1920x1080 with 16xAF and vsync off. All the videos are recorded using dedicated PC with capture card, there is zero performance impact on the tested CPUs. CPUs tested: Core i3-530 [2C/4T, 2.93 GHz] hw-museum.cz/cpu/463/intel-core-i3-530 Phe...
GeForce GTX 580 (OC) vs Radeon HD 7950 Boost - GPU Duel (ep. 36)
Просмотров 495Год назад
GeForce GTX 580 OC and Radeon HD 7950 Boost detailed comparison. Pretty much everything you can think of is here - fps, frametimes, memory usage, power consumption, and much more. Video Cards tested: GeForce GTX 580 (Gigabyte SOC) hw-museum.cz/vga/335/gigabyte-geforce-gtx-580-oc Radeon HD 7950 Boost (Sapphire) hw-museum.cz/vga/329/sapphire-radeon-hd-7950-boost Content: 0:00 Introduction 0:28 Di...
HDD Spin Up / Spin Down Sound
Просмотров 1 тыс.2 года назад
Recorded sounds of various hard disks from my collection. Starting around 1990 and up to 2008. The drives are in alphabetical order by the manufacturer and then by the year of introduction. Content: 0:05 Conner CP30174E (170 MB / 3822 rpm) 0:36 Hitachi HDS724680PLA380 (80 GB / 7200 rpm) 1:01 IBM DCAS-34330 (4 GB / 5400 rpm) 1:32 IBM DDRS-34560 (4.5 GB / 7200 rpm) 2:07 IBM DDRS-39130 (9.1 GB / 7...
Geforce4 MX 460 - Playing GTA3, Half-Life 2, Serious Sam: SE #GPUJune2
Просмотров 1,7 тыс.2 года назад
Playing the GTA 3, Half-Life 2 and Serious Sam: SE on GeForce4 MX 460. This GPU is limited to DX7 HW support, still it can play these games quite well. To get a bit of extra performance I decided to overclock the card. 375 MHz is a maximum for the GPU. The memory could probably go higher, but it is unnecessary as memory bandwidth is not the limiting factor here. This footage was captured using ...
GeForce GTX 560 Ti (OC) vs Radeon HD 6970 1GB (OC) - GPU Duel (ep. 35)
Просмотров 1,6 тыс.2 года назад
GeForce GTX 560 Ti (OC) vs Radeon HD 6970 1GB (OC) - GPU Duel (ep. 35)
Radeon R9 270X vs GeForce GTX 760 - GPU Duel (ep. 34)
Просмотров 2,7 тыс.2 года назад
Radeon R9 270X vs GeForce GTX 760 - GPU Duel (ep. 34)
Radeon DDR VIVO vs GeForce2 GTS - GPU Duel (ep. 33)
Просмотров 6642 года назад
Radeon DDR VIVO vs GeForce2 GTS - GPU Duel (ep. 33)
GeForce2 MX vs Radeon SDR - GPU Duel (ep. 32)
Просмотров 6633 года назад
GeForce2 MX vs Radeon SDR - GPU Duel (ep. 32)
Radeon HD 5870 vs GeForce GTX 480 - GPU Duel (ep. 31)
Просмотров 1,3 тыс.3 года назад
Radeon HD 5870 vs GeForce GTX 480 - GPU Duel (ep. 31)
Radeon HD 4870 vs GeForce GTX 260 - GPU Duel (ep. 30)
Просмотров 1 тыс.3 года назад
Radeon HD 4870 vs GeForce GTX 260 - GPU Duel (ep. 30)
GeForce 8800 GTX vs FireGL V8650 OC - GPU Duel (ep. 29)
Просмотров 4633 года назад
GeForce 8800 GTX vs FireGL V8650 OC - GPU Duel (ep. 29)
GeForce GTX 480 vs Radeon HD 5970 - GPU Duel (ep. 28)
Просмотров 9413 года назад
GeForce GTX 480 vs Radeon HD 5970 - GPU Duel (ep. 28)
Radeon HD 4870 (1 GB) vs GeForce 8800 Ultra OC - GPU Duel (ep. 27)
Просмотров 1,5 тыс.3 года назад
Radeon HD 4870 (1 GB) vs GeForce 8800 Ultra OC - GPU Duel (ep. 27)
GeForce 9600 GT vs Radeon HD 3850 - GPU Duel (ep. 26)
Просмотров 2,1 тыс.3 года назад
GeForce 9600 GT vs Radeon HD 3850 - GPU Duel (ep. 26)
Rally Championship 2000 Gameplay (Part 1) - Voodoo4 4500 PCI
Просмотров 9363 года назад
Rally Championship 2000 Gameplay (Part 1) - Voodoo4 4500 PCI
GeForce GTS 450 vs Radeon HD 5770 - GPU Duel (ep. 25)
Просмотров 9 тыс.3 года назад
GeForce GTS 450 vs Radeon HD 5770 - GPU Duel (ep. 25)
Mafia: The City of Lost Heaven Gameplay (Part 1) - Radeon 8500
Просмотров 4173 года назад
Mafia: The City of Lost Heaven Gameplay (Part 1) - Radeon 8500
Official Formula 1 Racing (Suzuka) Gameplay - Voodoo2 SLI
Просмотров 3,4 тыс.4 года назад
Official Formula 1 Racing (Suzuka) Gameplay - Voodoo2 SLI
NFS: Porsche 2000 Gameplay - Voodoo5 5500 AGP
Просмотров 7474 года назад
NFS: Porsche 2000 Gameplay - Voodoo5 5500 AGP
Geforce 2 MX = $119 Radeon SDR = $150
So $240 vs. $179, why not the mx460 against the 7500LE? Why not the $199 GeForce TI 4200? amd peasant fanboy much?
bro, the phenom x4 9650 supports DDR 2 1066.
It does, but in 4x2GB configuration it is not so easy and with this board such config wasn't stable no matter what I tried.
Зачем две HD2900? Она и одна в 60% игр быстрtе чем 8800 GTS 512mb
Back then you were less likely to come by the 8600gts in 512mb, most were the 256mb model. But the 2600xt did seem more common in 512mb. The 512mb version only did slightly better so probably was worth the extra cost back then. I saved my money for another month and got a 8800gts 320mb. It was one of my first high end cards.
For better compasion you need to use 1066 ddr2. This gives about 7 fps because Phenoms are very dependent on the RAM frequency.
Yeah, I would like to... but either the board or Phenom IMC didn't like my 4x2GB DDR2 memory sticks. It wasn't stable at 1066 MHz no matter what.
@@hardwaremuseum3532 Oh yes, thats a real challenge! I test about a 9000 RAMs with my Phenom x6, and now i have a 2 sets: 4x4=16gb Jezzalan 800, and 4x2=8gb HyperX 1066.
Very rare cpu battle!
i hope gamers nexus uses your video the next time they take shots at nvidia :))
in fact, Athlon 5350 is like Dual Core CPU with HT, those AMD architectures are like FX, not real Cores,
You say you overclocked the 560 Ti to 1 GHz, but these cards had a seperate clock for the shader domain. Did you OC this as well? It's interesting that the OC'd 560 Ti can match a 480, but power draw is also in the same ballpark - on paper, there is a 70W TDP difference between these cards. By the way, worth mentioning that GF11x had slight architectural and power efficiency improvements.
On all Fermi cards shaders always run at double frequency - in this case 2 GHz. As far as I know the ratio is set in stone, impossible to change (unlike Tesla arch.). Yes, there are some improvements. To be honest, it surprised me how well GTX 480 holds up. Fermi is supposed to be bottlenecked by low TMU count (GF100/110) and despite huge theoretical advantage of OC GTX 560 Ti in this area, the extra performance is not there.
@@MrA-ir3me Didn't know about the shader clock on Fermi, or probably forgot. Was indeed thinking of Tesla where it can be over 2x even at stock. As for Crysis 3, if my math checks out, this OC'd 560 Ti should have a theoretical spec of about 1536 GFLOPS, while the stock 480 is at 1344. That's about the same ratio. Perhaps that game is extremely shader bound?
I think it's cool that you have an Intel Pentium E5700 CPU basically a 3.0Ghz single core running at 3.7Ghz but it would have been nice to also include the fastest single core AMD X64 CPU like the AMD Athlon 64 FX-57 or something close to it just to compare retro CPU performance.
I am doing CPU benchmarks too. You can find them in my other playlists. Or on my website as articles. hw-museum.cz/
Попробуй ddr 1866 мгц,. CPU-NB на 2400 -2600 мгц, и нормальную видюху типа 1070, 1080 и сравнить его с четырёхъядерным Intel-Core i-7 сокета 1156 или 1366 с аналогичной памятью, а также оба процессора в разгоне (phenom x6 на 4Ггц). Нефиг шестиядерный феном тестировать со старыми видеокартами, с медленной памятью, медленным мостом и шиной по-сравнению с интеловскими, у которых шина и мост работает с завода на более высокой частоте.
What is wrong with serious Sam? Only 20 FPS but 100% cpu on 6600gt but 40FPS and 50% on 9800? How does that make sense? Did you mix it the two results by accident? 100% cpu on Radeon would make sense but this…
This is no accident because R9800 XT performs as expected when compared to other Radeons - 9700 Pro, X800 GT, etc. GeForce 6600 is rather slow here, I agree. Most likely this is because of low raw performance - both fillrate and memory bandwidth. Could be also some sort of driver bug. For more results you can check this article hw-museum.cz/article/13/the-ultimate-gpu-benchmark--2001---2005-/19
Hi HM! I know this may not be the right video to ask this, but I just installed NFS Porsche 2000 in my windows 10 PC. It is running properly and i did the 356, 356A, and 356B tournaments in EVOLUTION category. Now i am yet to do the 2000cc challenge. However my "Porsche Centre" is stuck to 2 cars only, the 911 coupe and the 356B GS Carrera 2. I am not able to see any other cars in the "CAR" dropdown tab. Is this a bug?
No
This fcking core 2 have More fps in cod Waw than my intel i5 1035g1 laptop processor
How do you measure PSU power like that?
Corsair RMi series can be connected by USB to the motherboard and HWInfo64 picks it up as another group of sensors.
@@hardwaremuseum3532 I see thanks.
two radeon x1900 xtx crossfire connect to ati mach64gt processor intel core 2 quad qx6700 conroe install windows xp
c2d power
GeForce4 MX did not support pixel shaders and this is bad. There were games that did not launch; there were no pixel shaders on the video card.
You are right. And at the time it received a lot of critics for that. Surprisingly, GF4 MX is very popular for retro PCs these days - very affordable, good compatibility with older games and acceptable performance too. Simply something like GeForce2 Ultra Extreme. :)
That's pretty cool, Phenom is competitive clock for clock, too bad the original release had all those issues and the Core 2 update in 2008 was a tough blow. On that initial Black Mesa benchmark they are almost on parity, I did not expect to see that! Great job as always!
For some reason Black mesa is much more AMD friendly than any other of my testing games. Well, I should add - K10 friendly. Because Bulldozer is much worse as I found out recently.
we really need a museum for this content and these chips :D
I have one online, check the links in the video description :)
what motherboard you used for core 2 extreme?
It is Asus P5Q-E. P45 chipset with DDR2 memory.
@@hardwaremuseum3532 is it cheap or expensive? and does supórt overclokcing?
The cost greatly depends on luck. But it is very good 775 board, so price can be rather high. Overclocking is supported and works well. I managed to OC C2D E8400 to 5 GHz.
Not exactly a fair comparison putting a consumer grade budget CPU handicapped with DDR2 against an HEDT offering. Nobody would realistically have been picking between these two when they were new. No offense, but it would also help if you put some background music in your videos or had some kind of commentary.
Actually, this comparison is not as far off as it might seem. I7-920 launched in November 2008, Phenom II X4 940 in January 2009... both for very similar price. The 1366 platform is somewhat beefier, yes. But LGA 1156 i7 or even i5 performs pretty much the same, at least with single GPU. The X4 940 uses DDR2 because it doesn't support anything better. And even if it did - I measured 5% performance gain on PII X4 965 DDR3 vs DDR2. I don't intend to add any commentary as I find it silly to just comment obvious stuff you can see on the screen. And as for music, it is way better for anyone to play their own favorite music. Personally, I always mute sound on PC hardware related videos :)
Given the comparative beast of a build otherwise, i expected these would be a lot closer to each other in performance.
This is why I buy AMD. They support their products for much longer than their competitors that abandon stuff after only a few years.
Thank you!
Those 2900 r bottlenecking by something, 2 x 38% Drivers motherboard chip or else i watch second comparsion w CF 2900 and its do wrong. I must complete my build and show up max potential od this cards :]
Coool!
did u enable the int gpu also with discrete gpu because dont forget that amd a10 has 12 compute cores 4 to cpu and 8 to gpu but they all are in same die and to get the juice of all 12 compute cores u need to enable int gpu as well with discrete.. This is the main reason of bottleneck that peoples dont enable int gpu from device manager after installing discrete card
In this test it was just the integrated graphics. I am aware of the possibility of crossfire with IGP + dedicated GPU, but I don't think it is a viable option.
I distinctly remember the FX 5800 being embarassingly slower than the Ti4600. did newer drivers solve all that?
The only case of FX 5800 being slower than GF4 Ti I can think of is when FX runs shaders 2.0 and GF4 runs shaders 1.3. Otherwise it is not possible as FX 5800 has much higher fillrate, memory bandwidth, everything. In apples-to-apples comparison NV30 will be always faster than NV25/28.
Why does the 800 card have a green tint on everything?
Most likely bad color calibration on the capture card. I didn't really check that when recording, only performance was interesting for me.
Dari dulu juga gitu, sjak jaman GeForce 4 atau 3. Kasta middle skrg kl ditambah Ti itu lbh bagus drpd kasta middle gen berikutnya. Coba aja 3ti vs 4mx Dulu masih penamaan MX.
never realize e7200 consume that low power. make me wana test my e7400
Maybe try to keep the memory the same.. the C2D has DDR2 speed advantage. The thing that stood out for me is how power efficient is the C2D. Amazing speed for the Watts.
Originally I intended to run the DDR2 at 1066 on all platforms. For some reason either the Phenom IMC or the board can't handle 4x2 GB DDR2 1066. Still, considering the slow FSB of the E7200 (only 1066 MHz), Phenom still gets better bandwidth even if running DDR2 800.
hey very nice video, is interesting how comparable this two are, and in some cases the athlon even matches the e7200. I'm pretty sure at the time if you couldn't afford the intel chip the amd was not that much of a bad choice
Like
new game please 🙏
Athlon 5350 4+4=8 core = Xbox one and PS4 CPU jaguar 😊
Actually, I have one of the chinese mATX boards with 8-core Jaguar. Perhaps I'll make a video with it some day. :)
@@hardwaremuseum3532 Oh, that would be great 👏👍
Best
The V8650 is somewhat a beefed up 2900XT or 2900XTX and it still fell short. The 8800gtx ultra really killed the 2900XT. If I remember right I got 2x 8800gts 320mb and my dad 2x HD3870 for around the same as a 8800gtx.
Yes, in this case I aimed for completely maxed out R600 - overclocked to 850 MHz and fast GDDR4 RAM. But as you say, still not good enough to match G80.
Amd cpu uses 40-70 watts more power so how is psu power only 20 watts higher. Is it because of cpu bottleneck GPU using less power with AMD cpu.
INTEL Toddler since when?
@@dallesamllhals9161?
Có bán cpu qx 9770 không
How interesting gtx one year older, less power consumption and still better than 2900
True competition to gtx 580 was 6970.
Wieso liegt die PSU Power beim i7 immer um die 100w höher ?
The X58 chipset is very power hungry, also the i7 has triple channel DDR3 while Core2 has only dual. And the most important - more load on the video card in case of the i7.
Great video as always
How about clock to clock comparison ? at 3ghz ? Iw will show IPC improvement Yorkfield vs Nehalem [like i7 750] without HT and with possible the same ram speed
Interesting you ask for that :) I actually did that few months ago as a part of my benchmark suite. Nehalem (3ch DDR3 1333) compared to Lynfield (2ch DDR3 1666) is very close -> 0 - 5% difference in favor of Nehalem. HT does nearly nothing in these games, only GTA V benefits realibly - 50 fps vs 45. Performance of Core2 vs Nehalem is very similar to results you can see in this video. Anyway - some numbers Yorkfield 12M vs Lynfield at 3 GHz: GTA V = 27 vs 44 fps, Dirt = 68 vs 101 fps, ROTTR = 34 vs 63 fps, Crysis = 46 vs 61 fps.
in Doom or Death Stranding difference beetwen those cards, would be even more. 780ti lose x2
Silent but great.
SLI Definitely not working, motherboard compatibility is probably an issue, as this card has requirements and will not work properly with many motherboards, probably need an NVIDIA chipset or cracked drivers.
Actually, SLI worked reasonably well - I doublechecked. The problem is 7950 GX2 is just that bad - and single chip 7950 GT even worse. There were some Nvidia cards that indeed needed NF4 SLI chipset - like 7800 GT Dual, 6800 GT Dual, 6600 GT Dual and such. All of those lacked PCI-E bridge onboard and required pci-e lanes bifurcation, which very few boards could do at the time. But 7950 GX2 has its own pci-e splitter onboard, so it runs fine on most chipsets and definitely is not tied to nForce.
@@hardwaremuseum3532 how did you check if it works? I remember there was an issue where the drivers and GPU would show it was working when it wasn't, and yes it was the motherboards issue.
Simply - GeForce 7950 GX2 is ~50% faster than single-chip 7950 GT. Therefore SLI must be working.