Huge FPS gains on PCIe 3.0 vs 2.0 @ 1080p in rFactor 2

[79wjd]

@LinusTech @Slick

 

I for one would like to see this, not to mention see the difference with the same system, except for an fm2 apu and a 4690k. Note I was running a gtx660 on both boards. (yes I know an apu is pointless with an Nvidia GPU, but the only dual graphics gpu for my apu wasn't as good as a 660)

Later today I could try running some benchmarks on my 4690k and 780 both on pcie 3.0 x16 and pcie 2.0 x16; it would be interesting to get some definitive answer. 

[totjup5]

Yeah, this is probably fake, since the bandwidth of pcie 2.0 isn't really a limitation there's no reason 3.0 would yield significant performance gains. This was either a flawed test, or an anomaly

[jimithebear]

I've found the two other system's DrR1pper was comparing his GTX 970 / i5-2500K @ 4.3GHz stats with;

 

1227mhz core/1753mhz mem
Min: 100 - Max: 149 - Avg: 126.399 (86.7% your Sentri's performance)


1316mhz core/1753mhz mem
Min: 103 - Max: 154 - Avg: 130.022 (87.8% your MJP's performance)


So you two are getting correct relative performance scores per your clock speed settings to one another but i'm losing out 12-13% performance. Hmmm

 

Sentri's running an I7 4790k 4Ghz. MJP - i5-3570K @ 3.6GHz.

 

Right, so technically we could dismiss the above comparison based on the differentiating hardware being used? Still doesn't explain MrPix's findings though?

[AMICLG]

Hate to sound noobish, but what's IPC? I'm just saying I did notice an extreme difference, I can now play things on high/ultra without severe choppiness, even with the same gpu.

 

Source?

IPC: it's basically how fast your processor can handle things, AMD CPUs have a much lower IPC than Intel ones.

 

The PCI-E generation isn't what causes the better performance but rather the better IPC on the newer processor being tested, the faster the IPC is the faster things can communicate meaning less bottlenecks. Bandwidth isn't the issue and has never been, PCI-E 1.0 @ x16 is equivalent to 2.0 @ x8 which is plenty of bandwidth but something that is old enough to have PCI-E 1.0 is going to have terrible IPC thus comes the bottleneck. Intel incresess IPC about 10% every tock for this reason though the performance difference after Ivy-Bridge is negligible only 2-3% realistically between Ivy and Haswell.

[The_Evenger]

i don't buy it.

[Sauron]

If you could hammer the card further (like enabling ShadowPlay for instance?) plus utilise one or two extra PCIe cards, could it spell more danger I wonder?

 

I highly doubt it. After all that would severely limit sli performance even in pcie 3.0 systems if pcie 2.0 16x/ 3.0 8x wasn't enough anymore. Nvidia and Amd wouldn't let that happen-

[jimithebear]

A very good point Sauron. So perhaps it's specific to the game itself as MrPix later wrote;

 

I have just tested in some DX11 games/simcades/sims and there is no such increase, in fact < 1% in most cases.

 

rF2 however seems to benefit greatly at the high end with 1080p resolutions and less so with higher resolutions.

 

Strange... a programming or compiling issue then? Is that even possible? Would be nice if a rFactor 2 dev would take a look.

[brownninja97]

Well either way the rfactor 2 community is under major flak as a result of assetto corsa

[deviant88]

A big load of BS... i ran my gtx 670 OC on my pci-e x4 2.0 slot and barely dropped frames in some games,only running afterburner osd is when i noticed a difference,but game still playable.

 

20-30% increase from x16 2.0 to x16 3.0 dream on baby,must be some bugging in the game or intentional code to make it worse,not even dual-gpu cards had problem with x16 2.0 over 3.0 in the past and neither would a 980,if that was the case even my 670 the 680/770/780/ti and amd equivalents would have reported the same performance hit.So far all pro reviews of various games shown likes less than 1-2% improvement from 2.0 to 3.0 since pci-e is still not a bottleneck,not even in theory.

[mvitkun]

oooooooooh, would that cause choppy framerates?

Trik all that @AdamIsaacLang is saying is that the Intel CPU has better per core performance.

IPC increased dramatically from i5/i7 -7xx/8xx/9xx (lynnfield, bloomfield, etc) to SandyBridge, to the point where, in certain applications that weren't threaded too heavily, the i5 2500K could outperform the i7 960.

 

Now ask yourself, "Does better performance per core, per clock, explain my getting better performance in games when the CPUs in question have the same number of cores or the Intel CPU has more cores?"

 

I mean you said you were surprised that you had better performance going from an APU to an i5 4690K, and that somehow surprised you?

APUs compete with the Pentiums, not even with the i3, so it shouldn't come as any sort of surprise that by 'just switching your CPU and motherboard' you'd experience a large performance increase.

[Kingofpants]

Did they plug the Monitor into the motherboard or something?

[Trik'Stari]

Trik all that @AdamIsaacLang is saying is that the Intel CPU has better per core performance.

 

Now ask yourself, "Does better performance per core, per clock, explain my getting better performance in games when the CPUs in question have the same number of cores or the Intel CPU has more cores?"

 

I mean you said you were surprised that you had better performance going from an APU to an i5 4690K, and that somehow surprised you?

APUs compete with the Pentiums, not even with the i3, so it shouldn't come as any sort of surprise that by 'just switching your CPU and motherboard' you'd experience a large performance increase.

I expected better, but not nearly as much as what I got.

 

That is, aside from my current problems with Nvidia's drivers for my gtx 660 >.<

[mvitkun]

I expected better, but not nearly as much as what I got.

That is, aside from my current problems with Nvidia's drivers for my gtx 660 >.<

What are the problems?

[The_Strict_Nein]

I noticed a complete and total difference when I moved from my fm2 pci-e 2.0 AMD board and APU to a lga1150 pci-e 3.0 board. Everyone tells me I'm crazy and that it doesn't make a difference, but I did notice one. Literally the only things that changed were the board and cpu, same graphics card, same monitor, same everything else.

 

You're probably noticing the CPU performance much more. Which CPU did you move to? Anything like an i3 or above >>>>>>>>> Anything on FM2

[Trik'Stari]

You're probably noticing the CPU performance much more. Which CPU did you move to? Anything like an i3 or above >>>>>>>>> Anything on FM2

I5 4690k, so yeah massive difference.

 

What are the problems?

Well, borderlands runs fine, but SWTOR just...ugh. The "driver shell" keeps crashing and recovering, this after the screen goes black, UI flickers, sound issues, ugh. It's just, constant problems. It's a FTW edition gtx660 which I'm being told has a lot of driver problems, what's even worse, it's less than a day old, just got it in the mail from EVGA, RMA'd my original one.

[The_Strict_Nein]

I5 4690k, so yeah massive difference.

 

Well, borderlands runs fine, but SWTOR just...ugh. The "driver shell" keeps crashing and recovering, this after the screen goes black, UI flickers, sound issues, ugh. It's just, constant problems. It's a FTW edition gtx660 which I'm being told has a lot of driver problems, what's even worse, it's less than a day old, just got it in the mail from EVGA, RMA'd my original one.

 

Yeah, you're definitely noticing the CPU, not the bus width.

[mvitkun]

I5 4690k, so yeah massive difference.

Well, borderlands runs fine, but SWTOR just...ugh. The "driver shell" keeps crashing and recovering, this after the screen goes black, UI flickers, sound issues, ugh. It's just, constant problems. It's a FTW edition gtx660 which I'm being told has a lot of driver problems, what's even worse, it's less than a day old, just got it in the mail from EVGA, RMA'd my original one.

You've already tried switching pcie slots?

[Trik'Stari]

You've already tried switching pcie slots?

Tried all 3 of the ones on my board. It's definitely the GPU or the drivers.

[mvitkun]

Tried all 3 of the ones on my board. It's definitely the GPU or the drivers.

Did you try older drivers? 

Did you have this problem while using the 660 with the FM2 build?

[Trik'Stari]

Did you try older drivers? 

Did you have this problem while using the 660 with the FM2 build?

1. No I did not have this issue, but I couldn't run things on high to begin with because it was just too choppy and crappy when I did try to.

2. I haven't tried older drivers, because I wouldn't know where to start.

3. It's not as bad as the issues I was having with my original 660, that worked fine for a few months, then started bsod'ing during any FPS games, but could run swtor on ultra with no probs what so ever.

 

This is a brand spanking new card, came in a nice plain EVGA box wrapped in plastic, not a refurb.

[Sintezza]

Its just bullshit lol.

Same story with corsairs high frequancy ram that time lol.

[deathjester]

Calling BS on this.

 

I5-2500k and a GTX 970 Shadows of Mordor. 73 FPS AVG. 377.75 FPS Highs. Doesn't look bottlenecked to me. Doesn't look like it needs a 3770k either, but we all knew that recommendation was horse@$^&. Lows could be better but that is API and he is recording footage on an I5.

 

[Common_Courtesy]

Yeah this sounds strange, I'd want to see more benchmarks from other people before believing this. 

[Cupitor]

new to the issue-

If such a drag exists, would this "effect" (what I have now)

be present /w a 3930k Sandy-E using an asus "PCIe-3" "ready" mobo?

Ty

[deathjester]

new to the issue-

If such a drag exists, would this "effect" (what I have now)

be present /w a 3930k Sandy-E using an asus "PCIe-3" "ready" mobo?

Ty

 

Looks like BS marketing to be honest. That 3930k and that platform is fine. More than fine. Spend the money on new GPU's or a 4k monitor or beer.

 

 

I know you can get "the itch", but those 2011 Sandy's still kick ass.