bottlenecking cpu

B

bdavis24

[H]ard|Gawd
Joined
Oct 31, 2002
Messages
1,667
i had a question.. my system is 5850x2 in crossfire 4gb ram 1.5tb raid 0 drives and amd 720be oc to 3.6ghz.. i ran 3dmark vantage on proformace setting and only got a score of 16000. isnt that low for my system.
my buddy has nvidia 260s in sli with a i7 oc to 3.7ghz and scores 22000.. how is his that much faster..
my only guess is my cpu is bottlenecking the hell out of my crossfire 5850s

anyone have any other thoughts
 
Phenom II 720 X3 at 3.6 GHz is not a bottleneck for 5850-X2 except in the rare situations (like World-in-Conflict) where a Quad core is necessary.

Ignore Vantage scores - it is useful for tracking changes in the same system - not for comparing performance with other systems
- i7 "scores" a lot higher - compare your frame rates in the Vantage mini-games with that of your buddy's PC
 
ok good idea plus i run a eyefinity setup and no game i have tried lags... even mw2 and battlefield 2 all features on 5760x1080. so all in all i guess it dont really matter much.. just was wondering if my cpu was bottlenecking my cards
 
The answer to this is very simple he has a nvidia card. On the vantage cpu physics test in nvidia cards assist the cpu with the test itself your ati cards do not. I ran into the same issue when i had a 3.2 quad core and 8800gtx. My cpu scored around 31000. Now when i upgraded to a 5870 I score less than half of that on the cpu score.
 
o wow never thought of that. so vantage is a nvidia benchmark i guess. was wondering how his cpu was scoring 30k
 
Yeah I thought the same thing after I upgraded to an ati. I believe he can disable it in the control panel which would give you a better look at how much closer the two systems are. Not positive though as I haven't had my nvidia in for a while.
 
bdavis24 said:
o wow never thought of that. so vantage is a nvidia benchmark i guess. was wondering how his cpu was scoring 30k
Click to expand...
That is why i suggested looking at the mini-game framerates

You also have to realize that Core i7 gets a LOT of points over Phenom X3
- Vantage is not a Nvidia benchmark although the CPU score is (over) inflated on the PhysX test
 
apoppin said:
Phenom II 720 X3 at 3.6 GHz is not a bottleneck for 5850-X2 except in the rare situations (like World-in-Conflict) where a Quad core is necessary.

Ignore Vantage scores - it is useful for tracking changes in the same system - not for comparing performance with other systems
- i7 "scores" a lot higher - compare your frame rates in the Vantage mini-games with that of your buddy's PC
Click to expand...

WIC is not quad optimized. I don't know why you think it is.

Slight better performance comes from cache not because it is quad optimized.
 
MrXROK said:
WIC is not quad optimized. I don't know why you think it is.

Slight better performance comes from cache not because it is quad optimized.
Click to expand...
WiC runs much faster on a Quad- vs tri- core vs dual-core - it is very clear from the minimums.

. . . i tested it. Clock for clock from stock to over 3.8 GHz
- with Q9550S vs. i7 920 vs. Ph II 955 X4 (playable) vs. 720 X3 vs. 550 X2 (unplayable) at 19x12 and 16x10 with maxed out settings
 
apoppin said:
Look at C2D vs. i7 at the same clocks - especially the minimums. i7 is faster clock for clock more than simple architecture

Look at the minimums here; it is far more apples to apples in the Phenom II line-up:

Core i7 vs. Penryn vs. Phenom II with HD 4870-X2 & TriFire
Click to expand...
I am just saying that having a dual core like the Core 2 Duo is actually better than having AMDs fastest quad in this and some other games.
 
Tell him to rerun it and cut off PhysX and or PPU in vantage options and then look at his score then. It will knock a good hunk out of it then. Plus him running it with PhysX on isn't a valid score compare.

There is also a new patch that fixes various issues and scores on vantage out also.

Vantage Latest version: 1.0.2
 
apoppin said:
Look at C2D vs. i7 at the same clocks - especially the minimums. i7 is faster clock for clock more than simple architecture

Look at the minimums here; it is far more apples to apples in the Phenom II line-up:

Core i7 vs. Penryn vs. Phenom II with HD 4870-X2 & TriFire
Click to expand...

You want to compare 2 different architectures?

Even your own tests show it's not quad optimized.

If you look @ Phenom x3 vs x4 at same clock rates the performance is exactly same up to 2 GPU cores.

Phenom x3 @ 3.9ghz 99, 50, 24
Phenom x4 @ 3.9ghz 99, 49, 25

Now if you look at tri fire you get slight better frame rates from quad. I don't know why that is other than tri fire needing extra CPU to run efficiently which has nothing to do with WIC being quad optimized.
 
apoppin said:
WiC runs much faster on a Quad- vs tri- core vs dual-core - it is very clear from the minimums.

. . . i tested it. Clock for clock from stock to over 3.8 GHz
- with Q9550S vs. i7 920 vs. Ph II 955 X4 (playable) vs. 720 X3 vs. 550 X2 (unplayable) at 19x12 and 16x10 with maxed out settings
Click to expand...

No not really. From a core 2 duo to a quad 2 duo the difference is cache in WIC. The quad has 2x as much cache that is why it's able to perform faster not because it's quad optimized.

Again same with x2 vs x3, x4. CACHE!!!!

x3 and x4 have same results since both CPU have same cache.
 
First of all, the Masstech engione for WiC is quad optimized:
http://www.firingsquad.com/news/newsarticle.asp?searchid=16276
We have been adding multi-core support to the engine to make sure it runs at its best on dual and quad-core machines.
Click to expand...

Another test:
http://www.tkarena.com/Articles/tab.../mid/382/ArticleID/53/PageID/238/Default.aspx
The official benchmark showed very similar results. The only major difference was at the minimum level. . . . The minimum fps on the dual core was 16, where as the quad core only went as low as 22. ...
Looking at the CPU graphs, there's no doubt that the dual core was working a lot harder than the quad core during this benchmark. Although the dual core was not working at 100% load, it was very close to it. The quad core on the other hand, looks like it was sitting back taking it easy, with plenty of power in reserve.
Click to expand...

More:
http://www.bjorn3d.com/read.php?cID=1162&pageID=3905
At the same frequency the quad core definitely beats the dual core CPU.
Click to expand...

http://www.yougamers.com/articles/4553_world_in_conflict_interview-page6/
It does scale, so a dual core runs faster than a single core, and a quad-core runs even faster.
Click to expand...

CLEARLY World in Conflict depends on multicore and not just cache :p
- look at the *minimums*
 
Last edited:
I am also wondering about a bottle necking CPU. I currently have a Q6600 (stock), 8gigs of ram, Win7, ATI 5870, and I play at 2560x1600. Was wondering if going to an i7 920 will make a huge difference in games like BFBC2 and ME2.

From all I have read, there just seems to be conflicting reports on real world vs benchmarks. What games, what res, etc.
 
Gmok Bonecrusha said:
I am also wondering about a bottle necking CPU. I currently have a Q6600 (stock), 8gigs of ram, Win7, ATI 5870, and I play at 2560x1600. Was wondering if going to an i7 920 will make a huge difference in games like BFBC2 and ME2.

From all I have read, there just seems to be conflicting reports on real world vs benchmarks. What games, what res, etc.
Click to expand...

For BFBC2 it definitely does, that game scales up to at least 12 cores / threads and the hyper-threading on my i7 has definitely shown a major advantage over other uses with Core 2 Quad CPUs.

But most likely some older multi-core games that had utilized up to 4 cores may only go that high and they won't make use for any more extra cores or hyper-threading. In that case the i7 should still give better performance due to it's general performance per clock speed ratio over the C2Quad series.
 
Derfman said:
For BFBC2 it definitely does, that game scales up to at least 12 cores / threads and the hyper-threading on my i7 has definitely shown a major advantage over other uses with Core 2 Quad CPUs.

But most likely some older multi-core games that had utilized up to 4 cores may only go that high and they won't make use for any more extra cores or hyper-threading. In that case the i7 should still give better performance due to it's general performance per clock speed ratio over the C2Quad series.
Click to expand...

Wow.Thank you.
 
Gmok Bonecrusha said:
I am also wondering about a bottle necking CPU. I currently have a Q6600 (stock), 8gigs of ram, Win7, ATI 5870, and I play at 2560x1600. Was wondering if going to an i7 920 will make a huge difference in games like BFBC2 and ME2.

From all I have read, there just seems to be conflicting reports on real world vs benchmarks. What games, what res, etc.
Click to expand...

You'd be better off overclocking the Q6600 for your money. i7 ain't worth it for gaming alone, and everyone knows it. Video editing, on the other hand..

Additionally with maxed-out settings you'll probably be bottlenecked by the 5870 before the CPU once you get to a certain threshold at that resolution.
 
MrXROK said:
You want to compare 2 different architectures?

Even your own tests show it's not quad optimized.

If you look @ Phenom x3 vs x4 at same clock rates the performance is exactly same up to 2 GPU cores.

Phenom x3 @ 3.9ghz 99, 50, 24
Phenom x4 @ 3.9ghz 99, 49, 25

Now if you look at tri fire you get slight better frame rates from quad. I don't know why that is other than tri fire needing extra CPU to run efficiently which has nothing to do with WIC being quad optimized.
Click to expand...

huh?
:confused:



Look at the *minimums*


Just compare Ph II 550 X2 vs. Ph II 955-X4 with both at 3.5 GHz at 19x12
4870-X2 = 14 vs. 21
4870-X3 = 15 vs. 22

Now compare Ph II 550 X2 vs. Ph II 955-X4 at 3.5 GHz at 16x10
4870-X2 = 14 vs. 22
4870-X3 = 16 vs. 22

Even with the details absolutely maxed out with 4870-X3 Tri-Fire plus 4xAA/16xAF - moving as much of the work to the GPU as is reasonably possible) - the Quad Ph II gets a much better minimum over the Dual-core Ph II at the same 3.5GHz.

The Tri-Core Ph II fits in-between and is probably the best bang for buck but still gives a bit less performance than the Quad Ph II at the same clock.
 
friend'scatdied said:
You'd be better off overclocking the Q6600 for your money. i7 ain't worth it for gaming alone, and everyone knows it. Video editing, on the other hand..

Additionally with maxed-out settings you'll probably be bottlenecked by the 5870 before the CPU once you get to a certain threshold at that resolution.
Click to expand...

Ugg. Two answers, and I am not sure which is correct. It IS strictly for gaming purposes, one in particular really. And it seems that BFBC2 benefits a ton from the CPU.
 
Gmok Bonecrusha said:
Ugg. Two answers, and I am not sure which is correct. It IS strictly for gaming purposes, one in particular really. And it seems that BFBC2 benefits a ton from the CPU.
Click to expand...

They're both pretty correct. The i7 will produce a small or very small (depending on the game) gain over a C2Q, clock-for-clock, so small that it's not worth the money for the full-blown upgrade unless you're also doing video work. In many games that are not threaded the Hyper-Threading will work against it and the C2Q might perform better. And in still other games, even with Hyper-Threading disabled the C2Q might eke out a win. Wait until the next Intel platform because you'll be wondering why you spent so much money on such little gain (for your specific application) otherwise.
 
friend'scatdied said:
They're both pretty correct. The i7 will produce a small or very small (depending on the game) gain over a C2Q, clock-for-clock, so small that it's not worth the money for the full-blown upgrade unless you're also doing video work. In many games that are not threaded the Hyper-Threading will work against it and the C2Q might perform better. And in still other games, even with Hyper-Threading disabled the C2Q might eke out a win. Wait until the next Intel platform because you'll be wondering why you spent so much money on such little gain (for your specific application) otherwise.
Click to expand...

Looks like I will hold out. Thank you all for the answers.
 
Gmok Bonecrusha said:
Ugg. Two answers, and I am not sure which is correct. It IS strictly for gaming purposes, one in particular really. And it seems that BFBC2 benefits a ton from the CPU.
Click to expand...

i7 is overrated for gaming by enthusiasts who have them. HT often interferes with frame rates in older games.

You are not bottlenecked with a single 5870 (as i was not bottlenecked with my 9550s or my8600 before it) as long as your CPU is overclocked to at 3.2GHz or so.

Take a look at the link i posted that pits i7 920 vs q9550s vs.Ph II x2, x3 & x4
- i used 4870-X3 tri-fire which is a bit faster than a single 5870

in the *future*, i7 will benefit games more
 
apoppin said:
in the *future*, i7 will benefit games more
Click to expand...

True. This will only happen as games are even more heavily threaded so that the Hyper-Threading can consistently be utilized effectively. But I have a feeling that by the time this happens to any appreciable extent, the next platforms will already be hot off the wafers --- and produce much larger gains. Then will be the time for the Q6600/Q9550 and other C2Q owners to upgrade.
 
I'm running a Q6700 @ 3.2ghz. From what I've seen and read, if you have a core2quad around those speeds, there is no reason to goto an i7 setup for gaming. It just simply doesn't offer much more.
 
apoppin said:
huh?
:confused:



Look at the *minimums*


Just compare Ph II 550 X2 vs. Ph II 955-X4 with both at 3.5 GHz at 19x12
4870-X2 = 14 vs. 21
4870-X3 = 15 vs. 22

Now compare Ph II 550 X2 vs. Ph II 955-X4 at 3.5 GHz at 16x10
4870-X2 = 14 vs. 22
4870-X3 = 16 vs. 22

Even with the details absolutely maxed out with 4870-X3 Tri-Fire plus 4xAA/16xAF - moving as much of the work to the GPU as is reasonably possible) - the Quad Ph II gets a much better minimum over the Dual-core Ph II at the same 3.5GHz.

The Tri-Core Ph II fits in-between and is probably the best bang for buck but still gives a bit less performance than the Quad Ph II at the same clock.
Click to expand...

If you are trying to find out if the game is quad optimized you compare it with CPU with same amount of CACHE!

X2 has half the cache as x3 and x4 CPU. If you compare x3 and x4 processors they have same results because these processors have exactly same amount of cache.

WIC is not quad optimized. The difference derive from CACHE.


Phenom x3 @ 3.9ghz minimum 24 fps
Phenom x4 @ 3.9ghz minimum 25 fps

Exactly the same performance whether it has more cores or not.
 
Any benches for WiC between a E8200 and a Q8400 to settle this matter once and for all? E8200 = 2 cores, 6MB cache shared between the pair. Q8400 = 4 cores, 2MB cache shared per pair (4MB total). Both have 2.66GHz clock, 333MHz FSB quad-pumped to 1333MHz, and are 45nm. As apples-to-apples as you can get for this silly argument.

If the E8200 wins, it isn't quad-optimized. If the Q8400 wins, it is.
 
The matter is already settled.

In this link, one of the game engine's developers stated they designed it to use up to 4 cores. It's in the paragraph underneath the screenshots of the game in action. Case closed.

Now I'm not saying it uses all 4 cores extensively or well, in fact most testing shows that it uses the additional cores lightly only. In the interview he says they don't use the extra cores very heavily because they wanted to game to scale down to P4-class single core processors.

I tend to believe the developers.
 
Crytek claimed Crysis was quad optimized but is it really? Tests show it's not quad optimized at all. Same could be said about WIC.
 
MrXROK said:
If you are trying to find out if the game is quad optimized you compare it with CPU with same amount of CACHE!

X2 has half the cache as x3 and x4 CPU. If you compare x3 and x4 processors they have same results because these processors have exactly same amount of cache.

WIC is not quad optimized. The difference derive from CACHE.
.
Click to expand...

Now you are being utterly ridiculous. i posted many links to many different tests besides my own - including quoting the devs - that indicate WiC is optimized for Quad core


it is clearly not cache :p
 
MrXROK said:
Crytek claimed Crysis was quad optimized but is it really? Tests show it's not quad optimized at all. Same could be said about WIC.
Click to expand...

Just because the framerate doesn't change with a quad core vs dual etc doesn't mean its not optimized. It could just be gpu bottlenecked, or there could be not enough workload to saturate all the cores, despite the work being divided up evenly.
 
apoppin said:
Now you are being utterly ridiculous. i posted many links to many different tests besides my own - including quoting the devs - that indicate WiC is optimized for Quad core


it is clearly not cache :p
Click to expand...

Your own test shows no difference from x3 and x4 cpu. :rolleyes:

Talk about ridiculous. You want to compare different architectures. When that didn't work you wanted to compare dual cores with half the cache as quad cores.

Developers claim lot of things. It's called marketing!
 
MrXROK said:
Your own test shows no difference from x3 and x4 cpu. :rolleyes:

Talk about ridiculous. You want to compare different architectures. When that didn't work you wanted to compare dual cores with half the cache as quad cores.

Developers claim lot of things. It's called marketing!
Click to expand...

Did you miss this?
:rolleyes:

encore:

First of all, the Masstech engione for WiC is quad optimized:
http://www.firingsquad.com/news/newsarticle.asp?searchid=16276
We have been adding multi-core support to the engine to make sure it runs at its best on dual and quad-core machines.
Click to expand...

Another test:
http://www.tkarena.com/Articles/tab.../mid/382/ArticleID/53/PageID/238/Default.aspx
The official benchmark showed very similar results. The only major difference was at the minimum level. . . . The minimum fps on the dual core was 16, where as the quad core only went as low as 22. ...
Looking at the CPU graphs, there's no doubt that the dual core was working a lot harder than the quad core during this benchmark. Although the dual core was not working at 100% load, it was very close to it. The quad core on the other hand, looks like it was sitting back taking it easy, with plenty of power in reserve.
Click to expand...

More:
http://www.bjorn3d.com/read.php?cID=1162&pageID=3905
At the same frequency the quad core definitely beats the dual core CPU.
Click to expand...

http://www.yougamers.com/articles/4553_world_in_conflict_interview-page6/
It does scale, so a dual core runs faster than a single core, and a quad-core runs even faster.
Click to expand...
 
You must log in or register to reply here.
Back
Top