Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/851
Article Thumbnail

Titan X Performance: PCI-E 3.0 x8 vs x16

Written on September 8, 2016 by Matt Bach
Share:

Introduction

A number of years ago, we published the article "Impact of PCI-E Speed on Gaming Performance" investigating whether you would see a difference between a video card in a PCI-E 3.0 x8 slot versus one in a x16 slot. At the time, we found that even the fastest video card available had nearly identical performance whether it was running at x8 or x16.

In the three years since that article was published, however, video cards have become much faster, higher resolution displays (including 4K surround) are more common, and video cards have become increasingly useful in professional applications. Because of these factors, we have decided to revisit this topic using the latest hardware to see whether there is now a measurable difference between PCI-E 3.0 x8 and x16.

Test Setup

For our test system, we used the following hardware:

To see whether there is a difference between PCI-E 3.0 x8 and x16, we tested with both a single GPU and dual GPU (in SLI where appropriate) with three gaming benchmarks and two professional applications - each of which are highly GPU dependent. These applications include Unigine Heaven Pro 4.0, Ashes of the Singularity, GRID Autosport, Davinci Resolve, and Octane Render. In order to thoroughly test the different PCI-E configurations you may run into, we tested single and dual Titan X configurations at full x16 speeds as well as limiting them to x8 by covering half the PCI-E contacts with a piece of insulating material (paper). The PCI-E configurations we tested are:

 
 
  • Single GPU in PCI-E 3.0 x16
  • Single GPU in PCI-E 3.0 x8
  • Dual GPU in PCI-E 3.0 x16/x16
  • Dual GPU in PCI-E 3.0 x16/x8
  • Dual GPU in PCI-E 3.0 x8/x8

Unigine Heaven Pro 4.0

Starting out our testing is Unigine Heaven Pro 4.0 which is a fairly standard gaming benchmark. It is beginning to be a little aged at this point, but it is still one of the best and most consistent DirectX 11 benchmark we know of. To test a range of different displays, we included results for 1080p, 4K, and 4K surround with a variety of quality settings.

Unigine Heaven 4.0 x8 vs x16 single GPU

With a single Titan X, there was little benefit (less than 2%) in most cases to running the card at x16 instead of x8. In fact, with a 1080p display, using PCI-E 3.0 x8 was actually about 9% faster than x16! This result is odd and not what we expected, but we ran the benchmark multiple times and got the same result over and over.

Unigine Heaven 4.0 x8 vs x16 SLI GPU
Moving on to the dual GPU results in SLI, we saw effectively no difference at 1080p and 4K, but we did see some significant differences when using 4K surround. At this extremely high resolution (11520x2160), we saw no difference at ultra quality settings, but as we turned down the settings (which resulted in a higher FPS), running the cards at x16 became more and more important. When we were getting around 45FPS, x16/x8 was still only about 1% slower than x/16/x16, but x8/x8 was almost 4% slower. Lowering the settings to the point that we were getting roughly 100FPS, we saw a 15% drop in performance with x16/x8, and a massive 30% drop in performance running at x8/x8.

Ashes of the Singularity

Ashes of the Singularity was one of the first DX12 games to include a built-in benchmarking utility and is especially interesting because the developers specifically say to take dual GPU configurations out of SLI and run them as two seperate cards due to how DX12 handles multi-GPU configurations

Ashes of the Singularity x8 vs x16 single GPU

With a single GPU, we saw fairly consistent results across the different resolutions. In all cases, running the card at x8 was about 2-5% slower than running at x16.

Ashes of the Singularity x8 vs x16 dual GPU
With two Titan X video cards, we had a little bit of trouble getting 4K surround to work properly - in fact, we were unable to successfully take the video cards out of SLI at all with surround enabled. Even physically removing the SLI bridge did not work as it simply made the cards give driver errors in device manager.

At 1080p, all the results were within 2% (and x8/x8 was faster than x16/x8) so we are inclined to call them all to be within our margin of error. At 4K, however, we saw about a 2.5-3.5% drop in performance with the cards running in x16/x8 and x8/x8.

GRID Autosport

GRID Autosport x8 vs x16 single GPU

With a single GPU, the results in this game were surprisingly consistent. Across the different resolutions we tested, GRID showed about a 5% drop in performance when running at x8 versus x16.

GRID Autosport x8 vs x16 SLI GPU
Unlike the single GPU results, when using two Titan X cards in SLI the results not nearly as consistent. At 1080p, running at x16/x8 was actually the fastest, beating x16/x16 by about 10%. Similarly, x8/x8 was also faster than x16/x16, but only by about 8%.

Increasing the resolution to 4K gave us results more in line with what you would expect. At this resolution, x16/x8 was about 2% slower than x16/x16 and x8/x8 was about 8.5% slower. Finally, with 4K surround we saw about a 5% drop in performance with x16/x8 and a 6.5% drop in performance with x8/x8.

Davinci Resolve

Davinci Resolve is a non-linear video editing and color correcting suite that makes heavy use of GPU acceleration for things like live video playback and exporting of video. While this makes it a great professional application to test PCI-E speeds, the one downside is that the benchmark file we are using (Standard Candle) to test playback speed is limited to 24 FPS and the built-in FPS counter can only give results in whole numbers. This means that while we may see some performance differences, we need to keep in mind that what appears to be a 5% difference in performance may actually be less due to the whole number limitation.

Davinci Resolve x8 vs x16 single GPU

The Standard Candle test file actually has 8 different Blur and TNR node tests, but since the simpler configurations gave a flat 24FPS, we decided to keep them out of our charts to help cut down on noise. Interestingly, while we saw a 1-2 FPS drop in performance with 30 Blur nodes and 4 TNR nodes, we did not see a drop in performance with 66 Blur nodes or 6 TNR nodes. This is likely due to the fact that Davinci Resolve only shows FPS in whole numbers, so we simply don't have a fine enough scale to see the difference.

Davinci Resolve x8 vs x16 dual GPU
With two Titan X video cards, we saw a small 1 FPS drop in performance with x8/x8 with 66 Blur nodes and 6 TNR nodes. However, with 4 TNR nodes we saw a 2 FPS drop with x16/x8 (roughly 8%) and a 3 FPS drop with x8/x8 (about 12%). While the whole number limitation keeps us from being super precise, it is still clear that there is indeed a benefit to using x16/x16 over 16/x8 or x8/x8 in Davinci Resolve.

Octane Render

Octane Render is one of the first GPU-based rendering engines to be developed and uses the GPU only (not the CPU) for rendering and has extremely good scaling across multiple GPUs. We've looked at x8 versus x16 performance in Octane only a few months ago in our Octane Render GPU Comparison article and while we saw no difference at that time, it is possible that the new Titan X will be fast enough to show a difference. One important thing to note is that full support for the new Pascal video cards (including the Titan X) is still not present in Octane. They work just fine and are still faster than the previous generation cards, but once CUDA 8 is released by NVIDIA there are expected to be a number of optimizations to even further increase performance.

Octane Render x8 vs x16 single GPU

With a single GPU, we interestingly saw about 2% faster performance with x8 over  x16. This is odd, but is actually almost identical to what we saw in our Octane Render GPU Comparison article linked above.

Octane Render x8 vs x16 dual GPU
Moving on to two Titan X cards, the results are all pretty much the same. There were small changes in performance, but nothing larger than 1% so they are well within our margin of error

Conclusion

Overall, the results of our testing is pretty mixed. With a single Titan X, we saw a wide range of results between using a PCI-E 3.0 slot at x8 and x16. Some applications (Unigine Heaven Pro and Octane Render) showed no difference, while others (Ashes of the Singularity, GRID Autosport, and Davinci Resolve) showed up to ~5% difference in performance. 

With dual GPUs, the results actually got a bit more confusing. Although Unigine Heaven Pro didn't see much of a difference with a single card, with two cards in SLI driving three 4K displays in surround we saw roughly a 15% drop in performance running at x16/x8 and a massive 30% drop in performance running at x8/x8. On the other hand, Ashes of the Singularity only showed minimal differences, and GRID Autosport was actually faster at 1080p when running in x8/x8 - although it was about 8% slower at 4K and 4K surround. On the professional side, Octane Render still didn't show a difference when using two cards but Davinci Resolve did see up to a ~10% drop in performance with both x16/8 and x8/x8.

While our results are somewhat inconsistent, there are still a couple of conclusions we can draw:

  1. Whether you will see lower performance with x8 versus x16 is going to highly depend on the application. Some may see a difference, others won't.
  2. The higher the load of the card(s) - either through a higher resolution in games or a large number of accelerated effects in professional applications - the higher the chance of there being a difference in performance

At this point, we would say that if you are using a high end video card like the Titan X (Pascal) or possibly even a GTX 1080, it is probably a good idea to try to use a PCI-E 3.0 x16 slot - especially in multi GPU configurations. Depending on the applications you use it may not make much of a difference, but the fact that we saw a 10-30% drop in performance with x8/x8 compared to x16/x16 in a couple of our tests just goes to show how large of a difference it can make in certain situations.

Tags: Titan X, PCIe, Performance
Nick Dedman

What about loading times for games? This will be the time that the PCIe bus is saturated the most, as during gameplay, most of the data is already in the memory.

Another example of when PCIe bandwidth can help is in low vram situations. I never used to be able to run the 4k textures in Shadow of Mordor with my 3gb 780 on a PCIe2 based system. It was far too jerky after a few mins of play. Upgrading to a PCIe3 based system (which also has over twice the system ram bandwidth) allows me to do this (and maintain over 60 FPS) with the same GPU/Vram for as long as I want.

This is because windows will swap out excess video data into the system ram, over the PCIe bus.

Posted on 2016-09-09 22:13:43
David Wujcik

Well, you're talking about loading times for games that are at best being read from a sata SSD at 500MB/sec... that'll only fill half a single pci-e 3.0 lane, and you have 15 others available =P I have flash storage in my box that supports up to 3GB read speeds and I have noticed no difference with game loading on that vs my slower sata drives.

Posted on 2016-09-13 18:09:01
Streetguru

I know you guys don't use AMD for whatever reason, but the Radeon Pro Duo would have made a lot of sense to test given that ya know, it's a dual GPU card and all

Posted on 2016-09-13 03:07:24
Shadownet

AMD is a whole different beast. They have allowed in the past PCIE 4X in triple card configurations, but considering this is a test of bandwidth, I would say that a similar result should be expected of their cards in said configurations, probably the biggest hit being running 3 cards at high resolutions.

Posted on 2016-09-13 03:18:29
Streetguru

Just would have been the most extreme situation possible, especially if they used 2 of them, plus AMD cards don't use a bridge anymore either, so that's even more saturation of the lanes

even X4 though shouldn't be too big a deal, at least for mid range cards maybe

Posted on 2016-09-13 03:26:33
Shadownet

I wish someone does the benchmark, you do bring up a great point. Perhaps with the next generation of AMD CPUs.

Posted on 2016-09-13 04:50:17
cat1092

Shadownet, I believe it's happened, AMD's latest Vega lineup doesn't outright beat the NVIDIA GTX 1080/1080 TI or Titan Xp, plus these 2,048 bit (with HBR2 or GDDR6 memory) AMD GPU's uses a whole lot of power while trying as hard as possible to win. With these cards & considering pricing before the mining craze took over, plus specs of these, should had beaten NVIDIA hands down w/out trying that hard.

Yet it didn't happen, no matter how hard AMD tried. They're in a better position to compete with Intel over CPU's than NVIDIA over GPU's. Especially given now that with updates/patches that'll 'fix' these Intel CPU's to have a 15-30% power loss, now has a chance to rewrite history to reclaim those 'edge' victories that Intel won year after year. Meaning that the Phenom ll X6 1100T likely beats the best Intel retail models they competed against, as long as the patches are installed on the Intel CPU's via OS updates. These 'fixes' aren't permanent, unlike firmware, rather placing a Band-Aid on a major wound that should require stitches. That's why the 'up to' 30% performance hit, depending on OS & CPU patched. Linux OS's will take less of a loss on performance, while heavy ones such as Windows 10 will see more of an impact.

Cat

Posted on 2018-03-04 00:46:01
smarine

I REALLY want to upgrade my current cpu, i5-3570,but there really is not much out there, that much better for the price. THis cpu stands strong still today-- the only way I feel I can get the bump I now want as my needs progress, is the intel $400 or so card that hits 4.7or so ghz .

My cpu is quickly becoming oudated given what I"m doing,but man to get much of a bump means spending big bucks that I"m not willling to atm--not yet. Sad huh --
There are some amd ryzen cpu's out there, but again I won't get MUCH of a bump unless I spend at least $300, and right now its just not worth it..my GPU is fine overall at gtx 940 2gb, and while hardly ideal or as fast as it gets, it does good, can't say I"m not happy at all with it.

As always money rules some of us must wait <sigh>
Anyway I came here curious about the pciE2 thing as I was thinking about going backwards for a moment and getting FX-8350, but too bad that its slower than what I have by too much, short maybe of mutli-core .

Hard to believe given the .6 lead FX-8350 haS over ghz but I guess thats how it is.

What makes my cpu faster when its .6 ghz slower on Single core, other than cache ?

How does that work..( if anyone is still here)

Posted on 2018-08-04 20:27:09
cat1092

smarine, I'm here!;-)

I'm still on Intel i7-4790K, i5-4960K, i7-4770 & AMD FX-8370, 8350 & one 6300 system(s) & for all of my unique needs, one or the other does the job. The best one (i7-4790K) is running a GTX 1070, easily capable of hitting a million points per week with the folding@home app. Most of the others, except one (FX-6300 on 2GB MSI GTX 960) are running a 6GB GTX 1060, either the EVGA SSC or FTW.

While I'm sure there will be gains with new GPU's, and as you said. 'money rules', am waiting until 2020 to perform a ThreadRipper build, by then should be 3rd or 4th gen, with 64 PCIe lanes, will be all that I'll ever need & then some.;-)

Am saving towards that day as I'm typing this post, will likely sell off a couple of the systems above just to pay for the GPU, will want a NVIDIA (EVGA) that's fresh on the market. That is, a few months afterwards, by then any driver issues will be solved & pricing stabilizing.

Fortunately, one thing in our favor, the mining craze is finally over, unless one has many expensive GPU's running 24/7, there's little money to be made. Most of those who are making the big bucks has a warehouse full of GPU's running with a staff to keep these blown off & ensuring all are running at 100%. It's too much of an undertaking for one to buy 8-10 GPU's now & get in, probably 3-4 years too late.

Initially, was going to do a Ryzen 1800X build & having so many PC's to upkeep, decided it would be best to wait until EOL for Windows 7 & go with an all out AMD TR build, more lanes than the Intel X99 platform & a lot more power as well. This will allow me to run a top line GPU, whatever that may be, plus two 512GB Samsung (current version) PRO NVMe SSD's in RAID for killer speed, plus a couple of WD Golds for data & other storage.

Backup drives are kept offline & unless in use, disconnected from computer to prevent possible encryption from Malware. While many may feel this cannot happen to us, anything's possible, all it takes is a single click. Especially if used my less computer savvy family members.

Your FX-8350 should be fine for at least another 5 years to come, possibly longer. It's the MB that'll likely wear out first & these now costs a small fortune, even used. So if you have several FX or AM3 Phenom CPU's around, may be a good idea to find a used MB, or if a friend/relative is giving one away, don't turn down. As the good ones are $200 or more for 970, $250 if the 990 chipset, these are used prices. Occasionally one may find a new AM3+ MB for under $150, but I can almost guarantee not much overclocking room.

Good Luck!

Cat

Posted on 2018-08-04 21:19:05
smarine

DO you have think , MB"s ( fx ones for 8350+) that have no pciE3 x16 will be a problem going fwd..atm my gpu isn't the fastest as I"m using gtx 950 2gb (I no slouch at all), but I wonder how future proof the pcie2 is ?

I"ve seen tutorials , even Tomshardware that show 8350 holding its own against faster cards than my 3570 ,,,honestly , its a shame the cpu's need so much power but that's the direction AMD went in.

My only concern with buying 8350 is the single core performance,and given I make heavy use ( doing game dev) of Unrealengine 4 which atm, only uses single core for most things it does , I"m a tad worried.

Rendering, light building ya, the 8350 would wipe intel,but most of what I do sadly , as relates to ue4 is single core.

ITs a bit of a mixed deal obviously ;)

I guess games are more and more optimizing for quad core+ so maybe overall the 8350 isn't such a bad deal even today ?

Posted on 2018-08-05 17:01:54
cat1092

It's not our fault that NVIDIA has outdone AMD every time, and if by any rare chance that they outdo the 1080/1070 releases, NVIDIA will simply release a 'Ti' version that'll be better. It's AMD that always places themselves in defensive mode, rather than offense. The year of 2008-09 may have been their last great one, putting a true quad core on the market for $99 (the Athlon II x4 620), along with the 630 (which I have) & 640 for a little more. That was the last time they've placed Intel on the defensive, 2-3 years earlier, they introduced dual core CPU's to compete, though many were underpowered (the Athlon X2 4040e-5050e series). Yes, I have a few of those also.

In fact, have a few AMD cards myself (the best being a MSI branded AMD Radeon 7770 OC edition, 1GB GDDR5), though anyone who has had both AMD & NVIDIA knows the facts. Pay less now for an AMD GPU that'll not outperform the NVIDIA equivalent, and pay more later when the electric bill comes in every month. If one's on the PC 8-12 hours per day, that extra $100-150 yearly will be party money lost.

I'm just telling the truth, own PC's of both brands. At least one of the 'FX' CPU's (that I don't own) requires liquid cooling, not even the best of Noctua air coolers will put out the fire. Which is really lost energy, if one thinks about it. Why run so much current in a chip to melt it, many would had preferred a stable chip that runs with air cooling. So much for a 'cheap' CPU. Throw in the cost of the cooler, there's an unlocked i5, minimum, plus the Noctua cooler.

If AMD would use honest marketing techniques (like stop calling a 3 & 4 core CPU's 6 & 8 ones), that would help their image. Also step out in front first, then do the talking. The folks at NVIDIA kept a very tight lid on the GTX 1080/1070 until just 2-3 weeks before release, not nearly a frigging year. Actions are always louder than words, and for now, AMD is spouting words. Has those who purchased NVIDIA 980 cards in January/February known what was going to transpire, would had hung onto $1000 or more, purchased a GTX 1080, and had plenty of upgrade cash left over. Now they'll be lucky to unload these for the price of a GTX 1070 FTW card at best.

Plus AMD stopped development of proprietary Linux drivers for Ubuntu 16.04/Linux Mint 18 for not only old GPU's, also those unsold on retail shelves. Classy corporation indeed. NVIDIA has proprietary Linux drivers for older, middle age & the latest GPU's.

So AMD, who couldn't afford lost customers, gave up more with a growing Linux usershare. It was they (Linux users) who forced Microsoft Windows to finally dip below the 90% mark & it'll be they again when reduced to below 80% & Linus Torvalds will be flipping a bird at AMD execs, just as he did with NVIDIA ones in a packed courtroom years back. Only this time, he'll be less forgiving. He knows where the money trail is & knows that AMD knifed Linux users, beginning with Ubuntu 16.04/Linux Mint 18 & many other distros. Yet the open source community put aside their differences & came up with a solution w/out the first morsel of assistance from AMD.

Yet I'm glad that AMD exists, because w/out them, we'd pay more for out great NVIDIA cards that'll still run perfectly fine on PCIe 4.0 slots when these ships. Competition is good for everyone, plus those of all income levels deserves as much of a card as desired. For some, AMD may be the answer, or the stockpile of otherwise unsalable lightly used GTX 980/970's (for $170-300) beginning to pile high. Wow, what a difference a few months makes.;-)

Let's see what AMD is going to do, other than run commercials.

Cat

Posted on 2016-10-06 01:10:59
Shadownet

That's a very interesting result, thanks for the update, I was just reading the older one the other night. I guess finally bandwidth and games caught up on multiple cards. The lane configuration of two x16 also limits the boards and cpus, which makes it a very expensive proposition for the time being.

Posted on 2016-09-13 03:22:36
Wascally

How would a plx chip affect the result? For example, a x99 motherboard with 40 lane CPU with a plx chip set-up to give 16x/16x SLI but more than 40 lanes utilized (two gpus, multiple m.2/u.2's, etc.). Thanks!

Posted on 2016-09-14 05:12:19
godisafairytale

I would also be very curious to know this!

Posted on 2017-03-01 20:13:59

4k surround. I stopped right there to take a deep breath.
My wallet is empty anyway.

Posted on 2016-09-16 18:39:40
cat1092

Yes, mine too! The EVGA GTX 1070 FTW (on promo) was $439 alone. That's not counting the i7-4790K (though that was purchased a year ago as an upgrade for a XPS 8700, never panned out because there was only a single 4 pin CPU power port, as a result, the PC would freeze in place (no BSOD, nor logs to show why).

Am hoping that on the PCIe 3.0 x16 slot, I can have decent 4K on a 24" monitor with the 1070 FTW, prices are falling as 8K is developing. Since I'm not into gaming, a SLI setup wouldn't be beneficial to me, that cash will pay for a monitor.

So the two good things of PCIe 4.0 will be that the PCIe 3.0 cards will work fine, and would be able to run my M.2 SSD in a native x4 slot, rather than the 'extra' GPU slot to achieve 2,500MB/sec read performance. The writes are as advertised on my x4 slot, the reads are 900MB/sec off.

And before anyone says that these SSD's are for bragging rights only, maybe some are doing it for this, however there's real benefits in a SSD that runs at 3x the speed (writes) & 5x the speed (reads). The computer is going to me more responsive in everything we do.

Unfortunately, this will come too late for the Z97 (or Intel 4th gen) owners, still the GPU & M.2 SSD in the PCIe slot will be beneficial. Maybe enough to invest in the platform. For me, that would be a new CPU, MB & RAM, the rest, I have, those three items are 60-65% of the build, as I use premium quality cases, PSU's & more.

And I still have a 2nd Z97 build coming, too late, have already paid for the hardware, just awaiting an RMA replacement on a MB, an ASRock Z97 Extreme6 that's probably better than my ASUS Z97-Pro Gamer. The Ultra M.2 socket is already on the MB & this is one of the few that has it.

Still, I prefer the PCIe x4 (or x8) slot to hold down heat.;-)

Cat

Posted on 2016-10-06 00:07:21
Mark

Some of these abnormal results are consistent with what you would expect to see when a CPU bottleneck is present. Perhaps we should take that into consideration. If it is true then the 4k results are probably more representative than the 1080p results.

Posted on 2016-09-30 02:51:06
David Keller

Edit, forgot about the 10 core 6950K

Posted on 2016-10-29 19:01:50
Kougar

Excellent article, thank you! Looks like single-GPU x8/x16 width still won't matter before PCIe 4.0 becomes a thing.

Posted on 2016-09-30 06:31:55
cat1092

PCIe 4.0 sounds appealing, though for the most benefit, the GPU OEM's would have to raise the memory standard, to GDDR6 or higher. Since there was no GDDR4, it's safe to assume that the next revision will be GDDR7.

May as well make all PCIe slots on the board capable of a PCIe x4 slot also, running at x8 speed. This would be extremely beneficial to those running M.2 SSD's as a PCIe device, rather than on top of a hot motherboard.;-)

cat

Posted on 2016-10-04 20:15:35
Kougar

Some AMD cards like the x1950, 2600 and 2900 HD used GDDR4. It didn't offer the expected performance for the high power/cost and GDDR5 replaced it within a year. Also Samsung has already announced it has GDDR6 in the works for the 2018-2019 timeframe.

I do agree the largest benefit of PCIe 4.0 would be M.2 slots. The second largest would be Intel's DMI 3.0 interface, which has the same bandwidth as that very same M.2 slot...

Posted on 2016-10-05 21:25:17
Jay Jardin

Thanks a billion. Thanks to this I had a reason to buy a 40 lane processor. I am super happy with my machine.
https://www.facebook.com/jay.j...

Posted on 2016-10-16 19:14:46
David Keller

Awesome benchmark. I was searching for a benchmark for PCI-E speeds with Pascal GPUs. Interesting results. I'm not gonna spend $1000 on an X99 system to gain 1-2 fps to run games at 4K. If you're running 4K surround then it may be a good idea. I'm currently at PCI-E 3.0 x8/x8 with my z97 OC Formula, 4790K and SLI 1080s. With all the games lately that don't take advantage of SLI, I have been wondering, since my mother board has PCI-E On/Off switches. If I might get better performance to just disable the bottom PCI-E slot so that my top slot will run at x16 with SLI disabled. But according to the benchmarks, it doesn't seem like it would be worth the hassle.

Posted on 2016-10-29 18:54:04
Eugene Ostroukhov

I'm very confused about results.
It means that I do not need to upgrade my Z87Deluxe motherboard to use M.2 SSD disks?
I just need to use PCI-E adapter? >.<

Posted on 2016-12-15 09:04:45
Roger Skullestad

One problem with the test is that when using a different socket, with 4 cores which has less lanes and must run 16x8x - it also will mean higher clock per core, which also usually boost games a few fps - this wont be visible when only taking away the benefit of 16x16x, but not gaining the benefit of not going with a higher core-count with a lower clock.

Posted on 2017-01-09 15:06:42
clint

One thing which may skew the results is if you leave the motherboard on the default auto setting for choosing version of pcie ( eg 1, 2 or 3 ). Windows with often change default 3 to 2 if the lanes are not running at high capacity. However running the card at x8 probably gives a high lane usage capacity and the auto setting will keep it on pcie version 3 to maximise data transfer.

Posted on 2017-02-07 21:51:07
Marko

Thank you for following up on this! In my personal experience with crossfire r9 nano, switching from z97 to x79, both with 4 core cpus, the results were dramatically better in games. There is visually less stutter as well as higher fps. I think maybe it is more crucial to have the full 16 lanes per card in crossfire due to lack of a bridge or other variables such as drivers. Would you please do an AMD comparison?

Posted on 2017-03-19 08:48:06
Olivier

I have a litlle question regarding the compatibility.
Are NVidia PCI-e 3.0 graphic cards working well on 2.0 mother boards?
In the old article it was clear that it works... but now with recent GC, it is still the case?
Thank you.

Posted on 2017-04-04 07:33:11

PCI-Express has done a good job with being backward compatible in my experience, and running a PCI-E 3.0 x16 capable video card in a 2.0 x16 slot is comparable (in terms of bandwidth) to running it in a 3.0 x8 slot.

Posted on 2017-04-04 15:56:47

I should also mention, though, that a system old enough to have only PCI-E 2.0 may be slow enough in other areas (like the CPU) to limit performance even if the video card does fine in a 2.0 slot. Just something to consider :)

Posted on 2017-04-04 16:01:48
Olivier

Many thanks for your clear answer.

Posted on 2017-04-06 06:15:00
Tech TTT

Hello,

Can you please make the same test using Z270 platform ?

Hardware :

Z270

I7 7700K Oc@5Ghz or more

and test the following modes :

1- x8 x8 mode (to compare the results with the X99 platform and see how much a faster clocked CPU will make up for the loss of x16 x16 SLI lanes)
2- Single card x16
3- Single card x8

Please do this .

Posted on 2017-04-18 01:47:07
xX-illusion-Xx

Just wondering how x16 x16 sli was tested as the motherboard does not support it.

Posted on 2017-05-25 17:48:01

The Asus X99 Deluxe II/U3.1 supports dual x16 when using slots 1 and 4. Those slots do not get reduced to x8 unless you have three or more PCI-E cards in the system.

Posted on 2017-05-25 18:00:06
xX-illusion-Xx

Sorry you're right, according to the Asus site it does, (with a 40 lane CPU and M.2 and U2_2 unoccupied). I had previously been looking at the specs on the build motherboard link, (below), in the main thread which says, (according to the PCIE specs), that it doesn't support sli x16/x16. I should have properly researched this first

Link In Build <https: www.pugetsystems.com="" parts="" motherboard="" asus-x99-deluxe-u3-1-10954#specs="">

Posted on 2017-05-25 18:31:44
Ragnar Lothbrok

Now test applications that actually push the limits with GPU's that do the same. Like Wticher 3 with a GTX 1080 Ti SLI setup.

Posted on 2017-06-05 09:08:01
janon

Too many people have, as usual, taken the conclusion drawn here, combined with no real deep knowledge, and turned it into “web gospel”. They then link the same 2 or 3 vids that confirm the bias.

The fact is the conclusions here are highly suspect and, even where potentially accurate, the scenarios are a niche in a niche.

The “30% gains” (which themselves weren’t even across the board) were not only when using SLI, but *in surround*. Single panel 4K showed no gains right here in these charts!

The conclusion really should be *if* you are running *4k surround*, then in *some* workloads there *might* be an impact.

Unfortunately legions of semi clueless now authoritatively state “you gotta have 44 lanes for 4K SLI brah!!!” as gospel which remains BS as per a PILE of pro testing.

Posted on 2017-07-20 18:11:58
JRoyal

It seems to me that the difference between x8 and x16 is negligible. It looks like you can run x8/x8 @1080p with little to no real difference from x16/x16. And at 4k there is no difference.

Posted on 2017-08-16 11:12:51
Idkanythingatall

So, if I was to build a machine with:

-Ryzen or Threadripper CPU
-x370 chipset (max x8/x8)
-Blackmagic Design Decklink (PCIe 2.0 x4)
-a 1070 or 1080 Ti
-32 GB RAM
-Samsung 960 Pro

for a workstation handling
-DaVinci Resolve
-After Effects
-Blender

I shouldn't notice a huge difference from an Intel build capable of x16/x16?

Gaming is nice but I'm more worried about color grading, editing, and visual effects

I'm trying to balance CPU and GPU performance for different tasks without spending more than I have to. Ryzen seems excitig, but I'm disappointed it won't handle x16/x4, but after reading this I'm wondering if it would even matter.

Posted on 2017-08-26 18:15:22

If you are just doing a single GPU, you should be able to run it at full x16 whether you go with Intel or AMD. Almost every motherboard has at least a couple PCI-E slots that use lanes through the chipset rather than through the CPU. X370 only has PCI-E 2.0 through the chipset, but since the Decklink card is only PCI-E 2.0 that shouldn't be a problem. Just make sure to check the specs for whatever motherboard you are considering to make sure it has slots with the capabilities you need.

Between Intel Core, Ryzen, and Threadripper I think the right choice depends on your budget. If a Ryzen 1800X system is at the top of your budget, that would be a pretty good choice. Intel is generally faster and cheaper for AE, but Resolve might be slightly better with Ryzen (we haven't done Resolve testing yet, however, so that is a guess). If you can afford a bit more, however, I would definitely go with a Core i7 7820X system. It should be decently faster for Resolve, AE, and Blender but will also be better for random things due to the higher single-threaded performance. Threadripper I wouldn't really consider unless you are doing a ton of rendering in Blender with a CPU-based rendering engine. It would be decent in Resolve, but AE and Blender it isn't really optimal for.

Posted on 2017-08-28 17:03:17
roadkill612

Perhaps, but life is short and time is money. Who needs all this angst intel deliberately inflicts on purchasing decisions, which you amply give examples of?

Its byzantine & it doesn't end there. It results in "ghost lanes" -resources that mysteriously disappear when a seemingly unrelated system change is made.

With few expensive exceptions, YOU CANNOT HAVE RAID NVME ON AN INTEL W/ GPU. For that you need TR/Epyc. Mark my words, raid will be top of your wish list soon, so u r mad to preclude this option by opting for a few FPS for a few more months til code evolves to use amd's fresh tools.

Get an amd and its simple. You get ~everything the architecture is capable of in any power level - no; ifs, buts, traps, surprise last minute extras.

Posted on 2018-01-15 17:39:08
Étienne

Hello,
I am using a server (for gaming), HP ProLiant ML350 G6 with 2 processors Xeon 5675, and 12 go of ram per processor (so total of 24 gb of ram)
My graphic card is on a x16 slots but in the ML350 G6's specifications, it is written x8 speed
Is it a problem for a Nvidia 1050 4Go ?

Thank you very much

Posted on 2018-06-27 11:56:31
smarine

Prob. not ,as tests have show, some tests were better with x8.
The go line is a bit slower faik anyway, so thats going to make it even more likely your current setup is more than adequate, depending what you do and what you expect to do , in the future.
My i5-3570 w/ GTx 950 2gb is holding up fairly well.I can run UE4 with a large complex terrain and water materials and lots of geometry with enough perk that editing all of that is very performant.

The only way I'm ever upgrading is when this rig breaks and then its going to be painful as the upgrade is going to cost me, being worth the price, prob.intel as intel atm has the highest ghz range even over ryzen 7 . Amd does beat on price though.

Posted on 2018-08-04 20:51:58

Why on EARTH would you use SLI to benchmark pro video apps? Run this test and indeed any multiple card set ups again with the second card not conected with SLI which doesn''t work with Resolve or Adobe or Cinema or anything! Also check the link speed of the card in HWINFO as I think it gets limited to PCIe2 or even one sometimes.

Posted on 2018-12-05 16:06:30

I'm not sure what you are referring to. This is an older article before we really fleshed out our DaVince Resolve testing, but I can assure you we didn't have the cards in SLI for the Resolve tests. If it had been in SLI, we wouldn't have seen any performance gain with two cards at all since Resolve can't see the second card when they are in SLI. We actually get into that a bit in one of our newer articles looking at NVLink (which requires SLI to be enabled on Windows): https://www.pugetsystems.co...

Posted on 2018-12-05 16:26:30

Thanks for clarifying, as many articles did suggest you were testing with SLI connected.
Nvlink works in nuke on quadro cards for some time now, and also asked direct gpu to nvme writing, by passing the cpu. Which could be interesting to investigate.

Posted on 2018-12-05 17:05:59

I didn't realize that Nuke supported NVLink, and I can't seem to find any documentation about it either. Any chance you have a link to something about it?

NVLink is super interesting, but the problem is that it needs to be supported by the software in order to function. I know of a few application in the engineering/simulation space as well as a few in the 3D rendering space (V-Ray for sure), but outside of those and some scientific/HPC stuff it is really early on in adoption right now. There are not too many workstation-focused software developers that were willing to spend the time on development when you needed super expensive Quadro GP100/GV100 or Tesla cards since not too many of their users had that sort of budget. Now that it is present on the much more affordable RTX cards, however, we are really hoping it will take off.

Posted on 2018-12-05 17:21:29

I could be getting confused. There was a direct link between ioFX early on on nvme times 2013 I think and if you had a quadro you could go nuke>GPU>ioFX bypassing the CPU to improve render speeds. I think a lot of stuff had to be bespoke scripted. I am guessing that is a lot of the extra optimisation that can make the difference between an OK workstation and a smooth happy one. I have a lot of undocumented jumpers on my motherboard I am dying to see what are there for... Thanks again for all help; it is an invaluable resource when troubleshooting

Posted on 2018-12-07 09:40:33