**all gamers read this**dual/quad cpu usage

i read this article as i was on the fence about grabbing a q9550 to replace my e8400.....my e8400 is stayin put...

http://74.125.113.132/search?q=cache...&ct=clnk&gl=us


"The reality with Call of Duty 4 is simple ... it just doesn't care much about the processor used. And that accounts for any resolution."

"Say you game at 1600x1200'ish. With Call of Duty 4 it just doesn't matter if you own a Dual, Quad 60USD or 1400 USD processor."


http://i76.photobucket.com/albums/j11/bf2nut/cod4.jpg
http://i76.photobucket.com/albums/j11/bf2nut/cod4.jpg

"So from to to bottom we made sure we got the basis covered. I really wanted to insert a low-budget dual-core processor in the charts just for the simple fact you can observe the scaling of the faster processors. The slowest Dual-core processor used is the AMD Athlon X2 4850E. You can pick it up at roughly 60 USD and probably is the most green CPU in this test, for that fact alone I love it.

On the other side of the spectrum we see the 1400 USD Intel Core 2 Extreme QX9770. Anything in-between these two processors is the baseline we need to form an opinion about.

Now above in 1024x768 you'll notice the effect of the processor, after 1600x1200 the graphics card becomes more of a bottleneck."





thoughts?.....

But everyone has known that Cod4 is all about the GPU for a long while now. There really aren't a lot of games that truly do a lot better with a good CPU...

at high res the gpu is the only thing that matters as long as the CPU is adequate... this is the case with most games

COD4 runs on anything i mean even a toaster.

I think we've been thru this b4.

Leeghoofd had a test article somewhere about ghz and mb cache for optimal frames per second.
I think the skinny was:
1) 3.0ghz processor and 3mb cache to avoid bottlenecking and optimal minimal cache,
2) There is no advantage of quad (or i7 according to Nvidia) over dual cores for gaming (3dmark total score will be higher due to cpu BUT NOT frames per second).

I think I was reading somewhere too that 1gb gpu memory was preferred for 1920 x 1200 resolution.

Corsair had a decent article on how 4gb ram over 2gb ram increased the minimal and average frames per second for modern gaming.

COD4 and Unreal Tournament aren't great for frame per second hardware tests...most hardware will score high.

Crysis is the benchmark but there are plenty of other games that will give you a better rounded gauge of your system.

I might be wrong but I think what I remember above rings true...

Quote:

---

Originally Posted by Kylzer

COD4 runs on anything i mean even a toaster.

---

lol....

I may be wrong but having, gone from my e6600 to a q9550 seem to help, maybe having the same bus speed as the mobo, I had my 4890 with e6600 3dmark06 score was 11280, with the q9550 I am now pushing 16000

Kinda old and not as complete as I wanted to get it, but still getting to the point...

Quote:

---

Originally Posted by dum

I may be wrong but having, gone from my e6600 to a q9550 seem to help, maybe having the same bus speed as the mobo, I had my 4890 with e6600 3dmark06 score was 11280, with the q9550 I am now pushing 16000

---

If you factor out the "cpu score" the results will be much closer.
Also compare the frames per second in the procyon/forest/etc.

I thought it was common knowledge that MOST game engines are not multi-threaded, or threaded well if they are.

Quote:

---

Originally Posted by Mechromancer

I thought it was common knowledge that MOST game engines are not multi-threaded, or threaded well if they are.

---

there's that too... these are the same charts people who felt jipped for buying FX-55s instead of 4400+s whipped out for the first few months of dual-core. just remember it will all hopefully change. 'they' need to come up with some algorithms to split the load onto however many cores there are... 1, 2, 4, 32...

Old game engine...

Old news is old. Seriously dude, the linked article is a year and a half old; lol @ thread title, double asterisks and "all gamers read"? This stuff is just common knowledge by now. What they're demonstrating there isn't CPU scaling, it's GPU performance - you run into GPU limitations at about 1280 and the effect becomes more pronounced the further up you go. Before discounting more cores entirely, though, I'd have a look at the World in Conflict page - in games with complex engines, more cores does indeed make a significant difference.

I should also point out that the test config there was a single card - CPU horsepower can make a significant contribution to SLi and Crossfire scaling.

Quote:

---

Originally Posted by AndrewZorn

at high res the gpu is the only thing that matters as long as the CPU is adequate... this is the case with most games

---

Yeah, I ran Crysis on high 1600x1200 with 2xAA with a PD820 and an 8800GT, I did have 4GB of RAM though. :p:

There r CPU hvy games that will benefit from a triple or quad core just not that many (yet) like for example GTA4.

Although i think i once read that GTA4 only uses 3 cores cause its a poor console port.

this might change your mind:
http://www.guru3d.com/article/core-i...ance-review/15

Quote:

---

Originally Posted by Starscream

There r CPU hvy games that will benefit from a triple or quad core just not that many (yet) like for example GTA4.

Although i think i once read that GTA4 only uses 3 cores cause its a poor console port.

---

True, it is a terribly optimized game.....

I think I read somewhere that pretty much ALL games only benifit up to 3 cores. The 4th really adds nothing. Might be wrong, but I think the article compared a Phenom x3 to a Phenom X4 at the same speeds, and it made no difference to the games tested.

The number of cores can make a difference even on a single threaded game if you set the os threads to three cores and the game thread to the remaining core, thus giving the game any entire core to itself. Granted not many people are going to do this but it is possible to do.

I still tell people if you game and don't multitask to any considerable extent while gaming then a dual core is all you need. If you are more forward looking and / or multitask ( or do A/V work ) then stick with a quad. Really all there is to it.

The way I use my system I prefer my Q6600 over my E8400.

Sounds like another reason to invest in GPU's, rather then CPU's... :)

I am a 1920x1200 guy, and run a 3.81GHz on a Q6600.

CPU bottleneck has never been my issue.

The Last Remnant loves my CPU, so much in fact it uses 70-80% across all four cores! Square Enix did a pretty good job of multithreading with UnrealEngine3 especially considering it was ported over from Xbox360!

Quote:

---

Originally Posted by LagunaX

If you factor out the "cpu score" the results will be much closer.
Also compare the frames per second in the procyon/forest/etc.

---

so really getting everything, the same bus speed is really not hellping at all ? I know most games don't use more than one core and some use 2, I just thought 4500 jump was a lot, kinda of new the cpu score was helping, but I did not think that much?

Well i use a E8400 @ 4Ghz and last night I was playing the street fighert 4 and the CPU use never past more than 60%.
But for example: Talonman have a GTX295 and a quad-core and he never had a bottle neck, but a friend have a E6850 (stock) and a GTX295 and he get a huge bottle neck... So this shows that CPU is importan for "scalability" on SLI or CF....dual or single GPU.

I do know for running dual 295's, the CPU makes a huge difference in the performance of Vantage. Probably in games too, but I cant speak to that. :)

http://www.evga.com/forums/tm.asp?m=...1&key=�

They were just using a single GeForce 8800 Ultra, correct...

I think most CPU's can keep that GPU happy.