Spring IDF 2006 Conroe Preview: Intel Regains the Performance Crown
by Anand Lal Shimpi on March 7, 2006 3:58 PM EST- Posted in
- Trade Shows
F.E.A.R. Performance
F.E.A.R. gets its own page for a couple of reasons:
1) It's the only gaming benchmark that we're using that doesn't use an Intel provided demo. This is the same demo we use in our tests.
2) The integrated test tool reports Min, Avg and Max results, and three graphs take up more room than one.
We ran with all of the effects settings at Maximum and the graphics settings at Highest defaults. Updated: As we've described in our follow-up article, there was an issue with the original F.E.A.R. results that has since been fixed. The charts below have been updated.
First up - the average frame rate:
If you had any doubts about the results on the previous page, this one should convince you. Even when running a non-Intel created demo, Conroe offers a 20% performance advantage over the 2.8GHz Athlon 64 X2.
The advantage exists in both the minimum and maximum frame rates as well:
Facebook
Twitter
RSS
220 Comments
View All Comments
JarredWalton - Tuesday, March 7, 2006 - link
Just to list a few:Doom 3 test uses standard file (Demo1)
Far Cry uses standard maps
Battlefield 2 tests have been linked a few times (I created them; ask for a link if you can't be bothered to dig it up)
HL2 maps have been made available before; I don't know that anyone else has our current maps, though.
FEAR uses the built-in benchmark
DivX/Xvid encoding (Sum of All Fears Ch.9) can readily be found if you want
Winstones is used elsewhere and *could* be downloaded/puchased (it's no longer officially supported)
Sysmark is used elsewhere and can be downloaded/puchased
Worldbench is used elsewhere and can be downloaded/puchased
SpecViewPerf is publicly available
PCMark, 3DMark are publicly available
There are plenty of tests that can be corroborated. Do some of our tests use internal files? Sure - every site does that. Generally speaking, though, when configured with identical hardware our results are consistent with what's reported elsewhere. Let's not forget that most other results are just as proprietary as any of our results; that's the real problem.
StriderGT - Tuesday, March 7, 2006 - link
Will you get into the trouble of producing numbers for the A64 X2 2.8 that would resemble intels's benchmark? At leats where this is possible. I think it would create lots of cretive buzz...JarredWalton - Wednesday, March 8, 2006 - link
I'd reproduce it if I had such a chip. ;)Maybe someone else at AnandTech can help out. If I can get my hands on an FX-60 and appropriate other hardware, I could give it a shot. The only test we can directly verify is FEAR, of course - the others used Intel demos, so we'd only be able to show scores for our demos on other games.
Questar - Tuesday, March 7, 2006 - link
Thanks Jarred, I guess I could have explained myself better.
That's basically my point. Nobody's tests have been independently verified. If someone is going to throw out the testing that Anand did at IDF, then pretty much ALL testing Anand has ever done needs to be thrown out also.
I'm attempting to tell people that you can't pick and choose for arguments, there's too many smart people here to fall for that. Either Anand's testing is valid (I think it is), or it's not. Someone can't say that one day it's valid and anothr it's not just because it doesn't show the results they want it to.
Sunrise089 - Tuesday, March 7, 2006 - link
No, this is wrong. What Jarred posted means that if you or I set up a rig with the same hardware as AT uses, normally an FX-57 and 2x512meg of DDR-400 2-2-2-? for game benchmarks we could get the EXACT SAME RESULTS AS THEM, within a reasonable margin of error, so long as you use the same testing files they did, which according to Jarred are freely available. Just because most other sites have slightly different test rigs by no means suggests AT's tests are not independently verifiable. To state otherwise calls into question AT's integrity as an information source.Questar - Tuesday, March 7, 2006 - link
That's the same thing I said.Sunrise089 - Tuesday, March 7, 2006 - link
I see that now. I think I let my desire to defent AT outrun my need to carefully read the post I think I disagree with. Sorry.Questar - Tuesday, March 7, 2006 - link
Hey, no problem!StriderGT - Tuesday, March 7, 2006 - link
Anand did not configure the test at IDF. They reported the results...HUGE difference
StriderGT - Tuesday, March 7, 2006 - link
"Okay, show me one web site that has ever independently reproduced Anand's results.To the best of my knowlege, none of the files that Anand uses for benchmarking are available to the public."
Timeless piece of human intelligence:
Okay I will show you some berries and my closet full of benchmark files
PS If you do not trust anandtech's reviews how do you trust them reporting intel's?!?!?