Crysis CPU performance: E6850 vs QX6850 vs QX9650 vs Phenom X4

harpoon

Gawd
Joined
Jul 14, 2006
Messages
829
OK, good news first - Phenom overclocks to 3GHz on the B2 stepping - yay! :)

Now to the bad news - Performance is poor, 5 - 10% slower than C2D/C2Q. :(

http://news.expreview.com/2007-10-29/1193590532d6599.html

Specs:
C2D/C2Q using P35 mobo/ 8800GTX / WinXP / Forceware 169.01
Phenom X4 using RD790 mobo/ 8800GTX / WinXP / Forceware 169.01
score.png


I have to say this is pretty disappointing, having a heavily overclocked Phenom fail to match a stock C2D/C2Q. Consdering Penryns are routinely overclocking to 4GHz+ on air... things ain't looking too pretty for AMD.
 

harpoon

Gawd
Joined
Jul 14, 2006
Messages
829
These benchmarks are GPU-limited, don't you think?

Probably to a certain extent, yes. For AMD's sake, I hope it's not. Hey, it is a 8800GTX...

If these results are GPU bound, then the Phenom X4 result is even more disappointing, if it can't keep up in GPU limited situations what chance does it have in more CPU bound situations?
 

vista_blista

Limp Gawd
Joined
Jul 21, 2007
Messages
480
Memory speed and looser timings might account for some of the difference... 5.5 - 7 percent in total. And with AMD's delays, perhaps they haven't had time to optimize. But otherwise it looks pretty good to me, if we're going to base it on a single game. Apparently only Intel, Nvidia, Microsoft, and Logitech are listed as Crysis partners. So I'll wait for a few more benchmarks before basing my decision. To be honest, I haven't played a video game in years. :)
 

vista_blista

Limp Gawd
Joined
Jul 21, 2007
Messages
480
You know, looking at the results, suddenly they either look really rushed, or just plain faked. Especially after seeing the crazy volts on the Phenom. What's the deal running the RD790 bus at 200 compared to the P35 @ 333MHz? And looks like the demo is only utilizing one core. I'm going to wait for something a lot more substantive before jumping to any definitive conclusions.
 

harpoon

Gawd
Joined
Jul 14, 2006
Messages
829
You know, looking at the results, suddenly they either look really rushed, or just plain faked. Especially after seeing the crazy volts on the Phenom. What's the deal running the RD790 bus at 200 compared to the P35 @ 333MHz? And looks like the demo is only utilizing one core. I'm going to wait for something a lot more substantive before jumping to any definitive conclusions.

I would have prefered the article to be in English... ;) perhaps someone who can read Chinese can translate for us?
 

cerebrex

[H]ard|Gawd
Joined
Sep 29, 2004
Messages
1,571
You know, looking at the results, suddenly they either look really rushed, or just plain faked. Especially after seeing the crazy volts on the Phenom. What's the deal running the RD790 bus at 200 compared to the P35 @ 333MHz? And looks like the demo is only utilizing one core. I'm going to wait for something a lot more substantive before jumping to any definitive conclusions.

regardless... if an overclocked Phenom is losing to a stock C2D you're wasting your time arguing, as the C2D will light it up when it's overclocked.
 

Digital Viper-X-

[H]F Junkie
Joined
Dec 9, 2000
Messages
14,964
whats up with the crappy timings, I hope its not a cpu issue >_>, and the 375 mem speed too

btw if the game doens't benefit from QC, which obviously we can see, and C2 Arch is still lets say 5% faster then phenom(K10) then obviously a DC c2 would still be faster in game by the 5% that the QCs are ahead by.

plus we don't know what the max oc is on that chip, or the c2, so its a baseless assumption to say that 3Ghz is the max on a b2 if AMD will launch 2.8 and 3ghz chips soon.
 

jon67

Gawd
Joined
Oct 29, 2004
Messages
576
So the above benchmarks are definitely GPU-limited then, as the results for the 3GHz dual-core (E6850) are spot-on the quad-core results (even a little faster than the Phenom). The small difference between the Intel and AMD CPUs may be ascribed to different bus speeds, memory timings, chipsets or whatever. All we can tell is that all the CPUs tested are capable of maxing out an 8800GTX in crysis at some given (?) settings.
 

fluxion

Gawd
Joined
May 31, 2005
Messages
864
Probably to a certain extent, yes. For AMD's sake, I hope it's not. Hey, it is a 8800GTX...

If these results are GPU bound, then the Phenom X4 result is even more disappointing, if it can't keep up in GPU limited situations what chance does it have in more CPU bound situations?

id imagine its the reverse. HOPEFULLY, for AMD's sake, its GPU-bound. as then, differences like system/memory bus have a bigger effect on performance than processor speed.

if it is gpu-limited, there's still a chance that Phenom has alot more headroom to compete with Intel on a cpu-limited application.

when something is gpu-bound you can basically underclock your processor and get identical results (not taking memory/system bus into account)
 

RB30

Limp Gawd
Joined
Sep 16, 2007
Messages
480
I have been running Crysis demo today with my Opty 185 and 8800 Ultra in both 1280x1024 and 1600x1200 under high and very high graphic settings. As far as I can tell, Crysis demo is GPU-bound because my CPU usage stays around 30-40%.
 

Menelmarar

Supreme [H]ardness
Joined
Feb 15, 2001
Messages
5,472
id imagine its the reverse. HOPEFULLY, for AMD's sake, its GPU-bound. as then, differences like system/memory bus have a bigger effect on performance than processor speed.

if it is gpu-limited, there's still a chance that Phenom has alot more headroom to compete with Intel on a cpu-limited application.

when something is gpu-bound you can basically underclock your processor and get identical results (not taking memory/system bus into account)
...If it's system/memory bus that becomes more critical. Wouldn't AMD pwn, everyone says their integrated memory controller and memory bandwidth stomps all over Intel...?
 

kllrnohj

Supreme [H]ardness
Joined
Apr 1, 2003
Messages
6,845
...If it's system/memory bus that becomes more critical. Wouldn't AMD pwn, everyone says their integrated memory controller and memory bandwidth stomps all over Intel...?

Could it be the RAM dividers at stock speeds then? I know my 4200 at stock the RAM, even though it is set to DDR2 800, is actually running at ~368mhz instead of 400mhz. That might be enough of a difference to cause a drop in memory limited situations, yes/no? If you look at the CPU-Z screenshot for it you will see that the RAM is running at 375mhz, not 400mhz... And the RAM in the Intel's have tighter timings despite running at a higher frequency - I'm calling BS

I'm not going to put much faith into that "benchmark" result anyway - heck, it could even be the motherboard they used (or drivers for said motherboard), or the Intel's had faster/lower latency RAM. I'll wait for a final chip and a proper review ;)
 

krameriffic

2[H]4U
Joined
Mar 1, 2006
Messages
3,214
Was this test done in the Crysis demo? I remember reading that it doesn't support multithreading anyway.
 

RangerSVT

[H]ard|Gawd
Joined
Jun 7, 2007
Messages
1,989
So many excuses for AMD's Phenom, were they this forgiving when Intel wasn't quite competitive a couple years ago?


Hmmm:eek:
 

debello_64

Gawd
Joined
Oct 18, 2004
Messages
672
So many excuses for AMD's Phenom, were they this forgiving when Intel wasn't quite competitive a couple years ago?


Hmmm:eek:

Well if they took a 2.4 or 2.6 Barcalona and over clocked it to 3Ghz and it is with in 2 FPS of Intels QX series in a dollar to performance point of view that isnt to bad.

Also consider that Intels $250 dual core performs exactly the same as the $1000 quad, doesnt say much about Intel either. Also Intels 45nm performs exactly as their 65nm quad, lends me to believe that any benchmark taken from the demo is not a true performance test of the game. I say this because the developers have been claming how Crysis will fully use a quad core and it aint showing..
 
Top