Those who use PC3200 in your m68xx/74xx!!!!!

I use Centon PC3200 on my 7422 and PNY PC3200 on my 6811, both from CompUSA. Both have CAS of 3, but still very fast.

Whats that mean? I use OCZ PC3200 1 GB. BIOS will only let it run at CAS 3, even though its made to run at CAS 2.5.

It's not much of an issue in the first place. http://www.xtremesystems.org/forums/...ad.php?t=48634

If there is a price difference, getting memory with faster timings is good pretty much just for bragging rights on an A64 system. Timing have virtually no effect on real world performance of an A64 system. The latencies are already as low as currently physically possible anyway.

PC3200 has better performence for games and other memory intensive tasks. Now wether the program you use takes advantage of faster memory is questionable but there are noticable performence gains. Sometimes as much as 10%.

http://notebookanalysis.com/modules....ewtopic&t=1090

Heres a link describing performence affect between having a A64 notebook with dual channel vs. upgrading pc2700 to 3200.

PC3200 is more efficient running smaller tasks while Dual Channel performence is only noticable with 1MB blocks and up.

There were definitely speed differences between pc2700 and pc3200. Even if you use clockgen to oc the ddr333 ram to 400mhz, it would still be faster. You can use A64 tweaker to change the timings also. I wish i also had dual channel too..

I guess .25% is still an increase I don't see paying for it though. PC3200 speeds synthetic benchmarks by a visible margin, but real world performance, gaming or otherwise, is virtually a wash. If buying new RAM anyway and there is very little price difference between PC2700 and PC3200, 3200 would be a good bet. If you are not replacing both sticks anyway and there is any more than a $5 or $10 difference, it's a waste an a S754 system.

The link u gave is talking about CAS timings, not PC memory speeds. The author used the exact same RAM, then changes the CAS timings in BIOS. CAS is totally different than MB/Second. You should reread your article.

Do the game and software tests yourself with PC2700 and PC3200 you would see a 10-15% difference. Not in all games but some.

This misinformation sometimes given on this board is ridiculous.

I have done the gaming and software tests myself on one single system, with same software environment. There is 2-3% on someapps/games, none in others. You can also "mock" the tests yourself if you don't have access to the different modules using the same memory while running it at the different speeds. Results are the same, ranging from undetectable to miniscule. And I agree about the misinformation on boards being rediculous.

Here is a test using a 4000+ showing higher than my own personal results, but still very low real world improvements. The margins being larger than my own tests are likely because of the faster overall system making the smaller differences more apparent.

http://www.sharkyextreme.com/hardwar...691_3481836__5

In each of the gaming benchmarks, there is no more than a 6% difference in FPS between generic 3200 and generic 2700. Last time I checked, moving a game from 20 to 21.2FPS (6% increase) isn't going to make it any more playable. Nor is going from 100 to 106 FPS going to make it any smoother. Again, if the price difference is negligible, of course it makes more sense to go with the 3200. I won't argue that. But there is no sense in spending much on whay is going to amount to nothing unless you are looking for higher synthetic benchmark scores at whatever price.

And hes using a ATI Radeon X800 XT 256-MB video card to do testing. That card doesnt even need system memory compared to video with 128MB or 64MB like most notebooks have.

Even with that awesome video card there was still a 5 frame difference between the PC2700 and PC3200.

Its actually very impressive the PC3200 had that effect with an ATI Radeon X800 XT 256-MB.

Try it with a 64MB MR9600 the framerate difference would be larger than 5fps. For lower end notebooks its pretty significant.

You just showed your lack of knowledge about hardware with those statements. Heck, I guess I'll just take all the RAM out of my desktop if it isn't needed then

The tests I personally ran, which yielder much closer than 5-6% difference in those tests were done on a higher end desktop with a PCI-E 6800GT, and a system using an AGP TI4600. The peroformance difference between DDR400 and DDR333 was even less (as it should have been) on the machine with the GF4 Ti4600.

Here is the explanation as to why that statement was ludicrous. As the GPU quality decreases, the CPU becomes the bottleneck, not the RAM. With any current game, the RAM (up to at least a gig, 2gigs in many cases with newer games) is under 100% load thanks to the extremely complex textures used in todays applications, regardless of GPU. A less powerful GPU doesn't put any more stress on the RAM. It puts a ton more stress on the CPU though. With the exception of cards such as the new hypermemory cards, system RAM has zilch to do with the GPU. Certain things are slated to VRAM and certain things are slated to system RAM. The system RAM doesn't take over what the VRAM can't handle.

Nice try at a comeback but i already knew system ram doesnt run video games. But it does help increase performence with lower end notebooks. especially notebooks that need every performence increase possible. Besides like i had stated a couple posts above, video games dont use PC3200 as much as applications and XP do.

But you showed a link anyway with video games having as much as 5 FPS increase with PC3200 using a high end graphic card. Which is even better than i thought.

First, it's not a try at a comeback. "Comebacks" stop in the 5th grade for most. It was clarifying misnformation:

If 5fps is more than you thought, then why were you trying to argue the facts in the first place? That just backs up what I said to begin with. Not to mention, an x800 is hardly a high end video card. It was at the time that test was performed though. None the less, the better graphics solution just helps the 5fps be that large in the first place (as small as 5fps is at that high of a framerate).

I looked around a bit and didn't find any with those timings. Not sure if your answer is in the BIOS or the on-die memory controller. It's really not that important though.

ThunderPC: You should probably quit now. Remember that there are some people in the world who think SLI is worth the money. Those people will always tell you that spending even three times as much for the most marginal of gains is the best thing ever. It's because of them that nVidia revived the junk concept of SLI, that Alienware is still in business, that limited-use ringtones cost $2.00, and that Britney Spears was given a place in the Rock and Roll Hall of Fame. They are living proof that if you put anything in a box and the label "better" somewhere on it, you'll have an instant market at any price. You frequently see these sorts of people spending thousands of dollars at the body shop trying to trick out their stock Honda Civics to be just like the guys in Fast and Furious, then doing things like challenging a Jaguar to a race. No matter how many times they lose, they always have an excuse that somehow doesn't include the fact that a 4 cylinder family car will never beat a 10 cylinder sports car in any category but fuel economy without opening the hood, since that expensive yet retarded looking spoiler has been proven in synthetic benchmarks to give at least a .005% performance boost. You can't win. Being right has nothing to do with it.

According to CPU-Z, mine is running 2.5. BIOS shouldn't have anything to do with it anyway. The memory controller is on the CPU. You can use a software based tweaker to change all the timings and frequency if you want. Mine runs at 2.5 without any tweaking though (via SPD). I can't get mine to run at CAS2, but some have reported success with the OEM RAM. The only thing I can change on mine is the frequency and 2T to 1T.

@ Andevian Don't forget the stickers that add 15 HP alone!