When will Conroes memory use be tested???

[thestain]

Lets see some test with all the image quality settings set to max on Monitors that can display 1920X1200, the cpus tested are for people who are willing to pay through the nose.. and the purely cpu, without how it handles memory test, miss half of the real-world performance for the bleeding edge high performance system as it will be used.

Current review.. just started.. no sli.. we need to see this.. no high resolution, yes still missing.. and image quality settings are not maxed out..

Come on Anand, give us the top of the line Enthusiast review we know you can.. ok??

 

[JEDIYoda]

Topic Summary: All the benchmarks avoid testing this part of Conroe's performance

too funny!! Yeah right!! rofl

 

[thestain]

While Conroe might kick butt, what is needed are benchmark test that will use all the memory bandwidth a cpu can handle and Anandtech appears to be going out of its way to avoid such testing.

Name one test that has been done in this regard.

 

[BrownTown]

maybe thats because so few things in real life are so hugely dependant on bandwidth, 2 channels of DDR2-800 is more than most anything needs. RAM bandwidth is jsut one part of the equation, and its one we all know that AMD will win, unfortunately for AMD its about the only thing they win, and its not gonna swing too many benchmarks in their favor since Conroe has so many things going in its favor (not hte least of which is 4MB of L2 which means you need less bandwidth since you are more likely to hit hte cache).

 

[thestain]

Agreed, but.. still would like to see the all out sli everything set to max image quality to see what kind of difference, even if GPU will play a larger role, since same for both.. just want to see the results.. Am I asking for too much??

 

[HopJokey]

Originally posted by: thestain
Agreed, but.. still would like to see the all out sli everything set to max image quality to see what kind of difference, even if GPU will play a larger role, since same for both.. just want to see the results.. Am I asking for too much??

Just wait until Conroe is offically released for a full suite of benchmarks to be run. Anand and the rest of the press will have the parts in their labs for comprehensive testing.

 

[HopJokey]

Originally posted by: thestain
Lets see some test with all the image quality settings set to max on Monitors that can display 1920X1200, the cpus tested are for people who are willing to pay through the nose.. and the purely cpu, without how it handles memory test, miss half of the real-world performance for the bleeding edge high performance system as it will be used.

Current review.. just started.. no sli.. we need to see this.. no high resolution, yes still missing.. and image quality settings are not maxed out..

Come on Anand, give us the top of the line Enthusiast review we know you can.. ok??

Also, this article you are refering to is a preview not a review. Reading the article it states that Anand and co. had to build two separate rigs from parts they got at Computex and many parts they brought with them on their carry on flight bags. So they did the best they could with what they were working with it seems.

 

[dexvx]

Originally posted by: thestain
Lets see some test with all the image quality settings set to max on Monitors that can display 1920X1200, the cpus tested are for people who are willing to pay through the nose.. and the purely cpu, without how it handles memory test, miss half of the real-world performance for the bleeding edge high performance system as it will be used.

Current review.. just started.. no sli.. we need to see this.. no high resolution, yes still missing.. and image quality settings are not maxed out..

Come on Anand, give us the top of the line Enthusiast review we know you can.. ok??

Apparently you don't know too much about hardware.

1) Almost any modern application will NOT fit in the L2 cache. SuperPi 1M doesn't and FeaR sure as heck doesnt.

2) At 1920x1200 (or whatever that res is), its entirely video-card limited, more so if you put AA/AF.

 

[thestain]

I guess what I am looking to find out how much the new Core 2 Duo will improve my gaming performance when I already am using two top flight video cards, ati 1900XTX on one rig and the 7900 GTX on the other.. it has been my experience that sometimes the cpu and memory do make a difference.. from a minimum of say 15 to a minium of 35 would be substantial and worth looking into an upgrade, but.. if it was only three frames per second... forget it.

With that large cache size of 4 MB, some test that do not usually use more than 2-4 MB of memory at any one time are going to have significantly better performance, compared to those that use say 50 or 100 MB at one time, don't you think?

Just a thought, but what are the default settings being used on the benchmarks, I understand they are adustable and are they being tweaked to use less or to use more memory??

 

[JPH1121]

Actually, from everything I've heard, 1MB of L2 on an A64 games better than the equivalent chip with 512KB.

I'm no expert though...I've only really ever had an AXP 2000+ and a San Diego 3700+ so I can't specifically tell you the differences between the cache sizes

 

[F1shF4t]

Originally posted by: JPH1121
Actually, from everything I've heard, 1MB of L2 on an A64 games better than the equivalent chip with 512KB.

I'm no expert though...I've only really ever had an AXP 2000+ and a San Diego 3700+ so I can't specifically tell you the differences between the cache sizes

Yea there is like 2% - 5% diff at most between 1mg and 512k cache. Biggest diff hapens at the 128k to 256k jump.

Conroes cache will not make a huge diff unless
1) intel used the larger cache to mask the slow performance of FSB
2) The programa are specifically written to use the large ammount of cache.

 

[kknd1967]

the same OP started 3 similar threads about Conroe vs A64 or K8L

hey Anand, please also put DVD burning quality with pi/po scan in your Conroe vs FX62 benchmark as these are real life performance too Clearly this is not a reasonable request as the determining factor is more on the burner. Similarly for your request, the determining factor is more on your own video card at high res high quality settings, at least for the current games.