GTA V CPU benches - AMD gets hammered (again)

[2is]

I didnt think you said the PS4 was better hardware, but the image quality of PS4 in the images above is sub-par vs the PC even at 1080p. The PC lighting, Anti-Aliasing, Texture Quality, shadows etc etc are superior even at 1080p vs the PS4.

Except the images aren't at 1080p they're at 4k. I've looked up 1080p vs 1080p. You aren't going to convince me my eyes were lying to me by showing me 1080 vs 4k images and saying the same would be true with 1080 vs 1080.

 

[Dave2150]

I'd hope 4K would look better than 1080p... I'm referring to apples to apples comparison 1080 vs 1080

1080P is on the low end peasant side for PC gaming now.

A move to 1440P or 1600P at least would put you ahead of the console crowd and afford you alot more detail and clarity in the games you play.

 

[jpiniero]

1080P is on the low end peasant side for PC gaming now.

At least according to the Steam survey, about 1.73% have 1920x1200 as their primary resolution and only 1% have higher than that. So you are talking about a small minority of gamers could even theoretically play above 1080p.

 

[2is]

1080P is on the low end peasant side for PC gaming now.

A move to 1440P or 1600P at least would put you ahead of the console crowd and afford you alot more detail and clarity in the games you play.

I'm well aware of the benefits of a higher resolution monitor. That's like saying higher settings are higher and are better looking than lower settings which are lower. Which has nothing to do with what I'm saying.

I'm not complaining that 1080 looks worse than 1440. I'm saying 1080 on PC doesn't look any better and actually looks worse in many scenes than 1080 on the PS4 in THIS game.

4K > 1600p > 1440p > 1080p > 720p

I know this already and it isn't my the argument I'm making. At equal resolutions the game should still provide for higher fidelity on high settings than it's console counterparts, PS4 in particular, but it doesn't. I know I'm repeating myself but seems like a couple of you aren't understanding my gripe.

 

[skipsneeky2]

I would expect that the vast majority of those 768p are laptops.

Perhaps,that and people who have oem towers that often time came with 17-19'' 768p panels.I know 3 people who have Dell/Acer/Emachine panels from 2008-2009 and they are all 19'' 768p panels.My wife has the Emachine one hooked to her lap top.

Newer oems mostly are using 20'' 900p panels but i am willing to bet those older towers that came with 17-19'' 768p panels are still operating and people are aiming to try GTA5 with it.Many don't know better if their C2D and 9500gt can play it but damn if they wouldn't try.If those towers aren't working or have been upgraded,i am betting many are sticking to their old monitors as well.

 

[RussianSensation]

A $330 GPU for your $135 CPU (8320E). Makes sense.

This thread is a fail. 4.3Ghz 8370E is $119, and for only $20 extra you get AM3 motherboard. At the same store 5820K is $299, 5930K is $499 while the cheapest i5 bundle is $250 USD.

So let's see now, who cross-shops a $140 8370E CPU+mobo Bundle against $250 i5 CPU+mobo bundle? :sneaky:

In the real world you can take that $140 AM3 8370E bundle then overclock and beat an i3 into the ground in 90% of PC games. While at it, use the $110 saved from not getting an i5, add $20 and go from a $150 R9 280 to a $280 R9 290X.

Apples-to-apples comparison would mean in the real world it'll be an i5 OC + R9 280 against an 8370E OC + R9 290/970/290X. Guess which system will win in 90% of PC games? The second one.

1080P is on the low end peasant side for PC gaming now.

A move to 1440P or 1600P at least would put you ahead of the console crowd and afford you alot more detail and clarity in the games you play.

No. LG's 1080P OLED beats any LCD/LED/Plasma ever made. At 55"+OLED, this TV would smash a 27" 2560x1440 one. To imply that 1080P gaming is for "peasants" is laughable in this context. Also, if you gave me a 4K 19" PC gaming monitor, I would donate it (i.e., I don't want it for free). Just looking at resolution without considering the monitor size, the IQ, is meaningless. Try telling someone who PC games on a large 40-65" 1080P Plasma / LED that 1080P is for peasants and they'll laugh at the concept of gaming on a 24" PC gaming monitor.

What's next, you are going to tell us that watching movies on a 28" 4K monitor is more immersive than going to the movie theater?

 

[jji7skyline]

I still think that 1440p at 27" is currently the best resolution to game at, especially if you get one of those overclocking korean monitors with PLS.

The next step up will be 1440p Ultrawide monitors at around 34" but right now they're a little too expensive.

 

[escrow4]

Techspot's GTA V CPU bench is non-sensical, as is buying FX for a gaming box. Look at pclab's or gamegpu's benchmarks.

 

[BigDaveX]

Except the images aren't at 1080p they're at 4k. I've looked up 1080p vs 1080p. You aren't going to convince me my eyes were lying to me by showing me 1080 vs 4k images and saying the same would be true with 1080 vs 1080.

Well then, would you care to show us which images you've been looking at? Because a 4K image scaled down should just look like a 1080P image that has 4X SSAA applied; it would make little-to-no difference to the lighting, shadow details or shader effects, which are where the big differences between the PC and PS4/Xbone versions are.

 

[mohit9206]

Performance on dual core Pentiums need to be further examined. Those minimum frame rates look scary and if it dips into single digits often then GTA V would be near unplayable on Pentiums.
Would also like Anandtech to benchmark GTA V on Intel and AMD igpu's to see if playable performance can be obtained at medium settings at lower resolutions.
I remember Anandtech used to include benches on resolutions as low as 800*600 until 2005 so some benches at 1366x768 and 1600x900 medium settings would be very much appreciated with both igpus as well as entry level and mid range gpus.

I would love to know how my 4.6ghz 760k fairs.

Why don't you test it on your PC and report back how it performs?

I still think that 1440p at 27" is currently the best resolution to game at, especially if you get one of those overclocking korean monitors with PLS.

The next step up will be 1440p Ultrawide monitors at around 34" but right now they're a little too expensive.

I think 20" 1600*900 is the sweet spot.
Much better than those 15" 1280*1024 CRTs. My Dell E2014H is pretty good.

 

[skipsneeky2]

I think 20" 1600*900 is the sweet spot.
Much better than those 15" 1280*1024 CRTs. My Dell E2014H is pretty good.

Only thing better were the Dell 19'' 1600x1200 monitors.I had one and apparently those used Sony Trinitron tubes which i could believe as it had amazing picture quality.

For a pure gaming box,i would use one over just about anything 20'' or below.Unless it was IPS of course.

 

[monstercameron]

Performance on dual core Pentiums need to be further examined. Those minimum frame rates look scary and if it dips into single digits often then GTA V would be near unplayable on Pentiums.
Would also like Anandtech to benchmark GTA V on Intel and AMD igpu's to see if playable performance can be obtained at medium settings at lower resolutions.
I remember Anandtech used to include benches on resolutions as low as 800*600 until 2005 so some benches at 1366x768 and 1600x900 medium settings would be very much appreciated with both igpus as well as entry level and mid range gpus.

Why don't you test it on your PC and report back how it performs?



I think 20" 1600*900 is the sweet spot.
Much better than those 15" 1280*1024 CRTs. My Dell E2014H is pretty good.

https://m.youtube.com/watch?v=7wrh5dlHknc

Gta v on apu is literally one YouTube search away.

 

[Kuiva maa]

Techspot's GTA V CPU bench is non-sensical, as is buying FX for a gaming box. Look at pclab's or gamegpu's benchmarks.

Techspot bench isn't sensical but not for the reason you state.

FX 8350 (runs games at 4.0GHz) does 72fps in their main chart but a 9590 clocled at 4.0 (identical to 8350 that is) does only 67 in their head to head comparison with 4770k. They have been doing these shenanigans for years over there (first time i personally noticed this 2 years ago during their Metro LL review where they have misreported their i7 numbers)

, probably randomly filling in framerates without actually testing. They may be a cut above pclab.pl (mother of all invalid bench sites) but that is not a compliment at all.

 

[SlowSpyder]

You want your GPU at 99%. That means it's being fully utilized.

Yea, I know. Just stating my observation that in this game I am not CPU bound with a 7970 @ 19x12 resolution (2xMSAA). So, even though some consider AMD to be getting hammered in this game, I still have more CPU power than I can use most of the time.

 

[TheELF]

FX 8350 (runs games at 4.0GHz) does 72fps in their main chart but a 9590 clocled at 4.0 (identical to 8350 that is) does only 67

9590@4,7 in the main chart also only does 74fps.

AMD ships out the 9590 with a much higher tdp than the 8350,if you only turn down Ghz without adjusting Vcore then it's not crazy to think that the same motherboard would have bigger trouble keeping things up to speed with the more demanding chip.

 

[Yuriman]

Yea, I know. Just stating my observation that in this game I am not CPU bound with a 7970 @ 19x12 resolution (2xMSAA). So, even though some consider AMD to be getting hammered in this game, I still have more CPU power than I can use most of the time.

Right. At stock, in the OP, an FX-8350 was showing ~40 minimums and ~60 averages. Frankly, that's about what I shoot for in most games with my 7850 so I'd be content with an 8350's performance in this case.

I'm glad I opted for an i5 at the same price though. I remember seriously considering an FX chip.

 

[frozentundra123456]

Techspot bench isn't sensical but not for the reason you state.

FX 8350 (runs games at 4.0GHz) does 72fps in their main chart but a 9590 clocled at 4.0 (identical to 8350 that is) does only 67 in their head to head comparison with 4770k. They have been doing these shenanigans for years over there (first time i personally noticed this 2 years ago during their Metro LL review where they have misreported their i7 numbers)

, probably randomly filling in framerates without actually testing. They may be a cut above pclab.pl (mother of all invalid bench sites) but that is not a compliment at all.

The 8350 would have turbo enabled while the 9590 downclocked to 4.0 would not. Considering that and the error of measurement, and that the reported difference is only 8% the numbers look reasonable to me.

 

[SlowSpyder]

Right. At stock, in the OP, an FX-8350 was showing ~40 minimums and ~60 averages. Frankly, that's about what I shoot for in most games with my 7850 so I'd be content with an 8350's performance in this case.

I'm glad I opted for an i5 at the same price though. I remember seriously considering an FX chip.

Nothing wrong with an i5, probably the more well rounded choice. That's what I'm seeing typically, 60FPS but some drops to 40FPS in a spot here and there.

 

[Kuiva maa]

The 8350 would have turbo enabled while the 9590 downclocked to 4.0 would not. Considering that and the error of measurement, and that the reported difference is only 8% the numbers look reasonable to me.

The FX8350 will run GTA V at 4.0 stock (no thermal leeway available-in comparison a 8320 will run it at 3.7) . But even If it would use its 100Mhz multicore turbo, it is ludicrous to say that 100Mhz plus whatever error margin there is (it is nearly zero if they run the largely deterministic game benchmark) would amount for such a difference when a 600Mhz-1Ghz (depends on turbo behaviour) difference going from 8350 to 9590 nets just 2fps or less than 3%. It is just the usual botched bench techspot produces when they "examine" frequencies.

 

[frozentundra123456]

The FX8350 will run GTA V at 4.0 stock (no thermal leeway available-in comparison a 8320 will run it at 3.7) . But even If it would use its 100Mhz multicore turbo, it is ludicrous to say that 100Mhz plus whatever error margin there is (it is nearly zero if they run the largely deterministic game benchmark) would amount for such a difference when a 600Mhz-1Ghz (depends on turbo behaviour) difference going from 8350 to 9590 nets just 2fps or less than 3%. It is just the usual botched bench techspot produces when they "examine" frequencies.

I am not vouching for the benchmarking of that site in general. However, can you document that 8350 does not use turbo at all in the system they are using? If cooled properly, I should think it would use turbo, or at least you cant be certain that it did not under the conditions of the test. I also dont accept that the error of measurement is nearly zero, especially since we dont know how they ran the test. Do you think they did several runs under each condition and averaged the data? That is the only way you could assure even close to zero error of measurement. I have worked in research labs for years, and trust me, anything under 5% error of measurement is exceptional.

Granted the numbers dont add up perfectly, but I just dont see anything there that would merit calling "shennagians" on the site, especially since that implies some kind of intentional bias, not just somewhat sloppy measurement. But this is off topic and pointless. I dont see any point in continuing this argument.

 

[Kuiva maa]

Υeah, I can vouch for the processor. Any halfdecent gaming load and the stock 8350 just won't turbo, I've toyed with a few.The 8320 on the other hand does even BF4 at 3.7 and Intel processors often game even at 400Mhz above stock (depends on motherboard) without overclocking at all.

Just go through their older reviews. I do not think it is necessarily bias, mostly sloppy/bored journalism (I think they haven't ever even bothered to bench mantle, for example). Very few sites do a decent job benching these days.

 

[Blitzvogel]

Finally picked up my reserved retail copy. I'm in the process of making the final update download, but I'm glad the 7x DVD installation is over at this point.

It's good to know my 4690K will have no problem. I'm more tentative in regards to my R9 270. I'll be gaming at 1600 x 900, but preliminary benches seem to indicate that almost maxed 900p steady 30 FPS is a no go. I've waited a year and a half for this, so I'm ready regardless. Granted I might still just have to install the game on my fiancee's i3-2120 + Radeon 7750 machine for poops and laughs

I'm not at all surprised seeing HT Intel dual cores running the game, they are as powerful as the CPUs in the PS4 and Xbone, which is kinda sad, but it's just the current nature of console gaming economics.

 

[soccerballtux]

I agree about 720p,people like me game at 720/768p on televisions.I am moving up to IPS/1080P in a few weeks but still more reviews need to include 768p.

Steam stats show 768p being the second most used resolution period...that has to be worth something.4k and a 680 with very high settings getting 18fps,so enjoyable.:awe:

nobody in the market for graphics cards has a 720p TV

 

[Dave2150]

No. LG's 1080P OLED beats any LCD/LED/Plasma ever made. At 55"+OLED, this TV would smash a 27" 2560x1440 one. To imply that 1080P gaming is for "peasants" is laughable in this context. Also, if you gave me a 4K 19" PC gaming monitor, I would donate it (i.e., I don't want it for free). Just looking at resolution without considering the monitor size, the IQ, is meaningless. Try telling someone who PC games on a large 40-65" 1080P Plasma / LED that 1080P is for peasants and they'll laugh at the concept of gaming on a 24" PC gaming monitor.

What's next, you are going to tell us that watching movies on a 28" 4K monitor is more immersive than going to the movie theater?

This is probably down to your age.

Gaming on a 1080P on a huge screen would yield such low pixel density and lower imagine quality, most gamers would hate it.

Though for those with ageing eyes I can see how it may look great.

Also, if you're unable to distinguish the vastly superior imagine quality a 4K monitor offers over a 1080P monitor, then your eyes are definitely failing you.

1440P and above (including 4K) are all solid upgrades for gamers, if they have the GPU horsepower required to adequately drive a high FPS at these resolutions.

 

[skipsneeky2]

nobody in the market for graphics cards has a 720p TV

I would figure more people would be hooked up to 26-32'' t.vs which usually more times then not support 720/768p then there would be 4k users.

More reviews should include lower end cards at least catering to the resolutions those cards most likely can only support.My cousins 1gb gtx650 with the lowest of settings apparently hits 997mb according to the menu at 768p,willing to bet that card couldn't do 900p without stuttering.That would go for any 1gb card either its the 9800gt recommended or one of the rare 7870 1gb cards which happens to be the fastest 1gb card ever made as of today.