Cl2 versus 2.5 official article discussion here.

ONly an idiot would compare ddr latency performance on a p4, the p4 is designed to hide latency and favor bandwidth, the article once again should have been entitled x ram performance on the P4 PLATFORM.

The gains shown in this article were not indicitive of ddr performance but of the p4's use of ddr, at least run the test on an athlon processor for some comparitive results.

Lazy lazy lazy, bad thg.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:

Add a reply

Register or log in to remove.

I think they should also take into 1 more account when they do benchmarks. The people who have bought the high performance cl2 ddr ram can run in two different setups.
I thought I would have supreme performance with my Athlon xp 1800+ and corsair ddr 2400 cas 2 on a soyo dragon plus.
well when running everything at 133 mhz in turbo mode with most aggressive memory timings i got a certainset of scores.
but sandra also told me i could run my memory at 133mhz cas2 or up to 143 in cas 2.5, i now run my system at 143mhz with cas 2.5 and the performance is about 12% higher. some could be due to faster processor clock now, but it is also higher performance on the ram. so, i think when the do a reviw like this.. they should show the faster better memory running with a higher latency as well as a higher bus speed as a performance alternative too.

Another Happy Abit BP6 Customer.

Reply to Anonymous

I totally agree. Did you look at the difference in the Sandra scores in high speed and CL setting of 2.5 and the lower speed memory with CL 2 settings. They were the same. If they would have run this on an Athlon system the CL 2 setup would have probably posted better results than the CL 2.5 with the higher speed memory.

This article was a waste of their time and our's. When are they going to do an article solely on the basis of current architecture and not on a platform's use of that memory architecture. There has to be a way to fun C3, Alpha, AMD, Intel, and other chip platforms, servers and desktops, with current memories and come up with a conclusion on the effectiveness of the memory itself.

They might have well have done a comparison of EDO and BEDO with different timings setings.

"Sometimes you can't hear me because I'm talking in parenthesis" - Steven Wright :lol:

Reply to bum_jcrules

Well, the performance results on a P4 are important regardless of WHY they perform that way... provided you own or plan to own a P4. I don't think they should choose one over another, but rather they should perform the tests on both so that people know exactly what their getting, even if they ARE buying the RAM for a p4.

by the way, can anyone tell me why I can't get cl2 registered pc2100 from crucial.com? IS THERE such a thing as cl2 registered pc2100? it only offers me cl2.5. How does registered memory affect performance, anyway? (I need it for a dual athlonmp board.) Thanks.

As far as THG not giving all the info to their readers for these reviews, I've noticed that the same is true in other cases as well. For instance, in the most recent mb roundup for the via kt333 chipset, they never discuss how a 166mhz fsb on the chipset bottlenecks at the 133mhz processor. In fact, in all of their discussion of the chipset ANYWHERE they never mention this. I understand, however, that thg is read by a certain caliber of techie who would probably know that anyway. So, where should they draw the line? At what point does informative become pandering to the lowest common denominator? It seems to me like a simple reminder that "psst. Hey. If you're looking for better speed and you don't overclock, don't bother wasting hundreds on a new motherboard and ram." wouldn't be too much to ask.

Studies have shown that most people prefer the taste of AMD to the taste of Intel

Reply to shmeggegie

Yeah, maybe testing an AXP at 166MHZ FSB from now on, would really kick ass, and would show more performance squeezed.

--
I set my graphics aperture to 1024MB, why can't I open anything now? :mad:

Reply to eden

Quote :

It seems to me like a simple reminder that "psst. Hey. If you're looking for better speed and you don't overclock, don't bother wasting hundreds on a new motherboard and ram." wouldn't be too much to ask.

with this point of view, you wouldnt able to inquire for because this forum wouldnt exist as the THG site. if the last IT newest products dont interest you, why are you here?!?
maybe it would be more related to stay with your old rig & try another hobby. ll

thinking your asking could help, sometimes...

if you know you don't know, the way could be more easy ...

Reply to labdog

Quote :

Well, the performance results on a P4 are important regardless of WHY they perform that way... provided you own or plan to own a P4. I don't think they should choose one over another, but rather they should perform the tests on both so that people know exactly what their getting, even if they ARE buying the RAM for a p4.

You cant make an article stating once and for all why cl2 is better than 2.5 on ddr and use a system to run the test which for all intents and purposes ignores the effect of that cas latency benifit.

That would be like running a latency test on ddr ram using a 1 mb file on a processor with 2mb's of cache, [-peep-] POINTLESS.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:

Reply to Matisaro

Quote :

with this point of view, you wouldnt able to inquire for because this forum wouldnt exist as the THG site. if the last IT newest products dont interest you, why are you here?!?
maybe it would be more related to stay with your old rig & try another hobby. ll

thinking your asking could help, sometimes...

well, I'll try to make as best sense of your post as I can.
you seem to be saying that if Tom's hardware told people that ddr333 bottlenecks at the processor unless you overclock that this would in some way eliminate the community forums. Well, that's just not true. Further, my interest in the latest technology doesn't necessarily mean I have to overclock or go out and buy the newest mobo and RAM every single time a marginally better performance increase might be possible, especially since a newer better one is right around the corner (inevitably) anyway. Now, I appreciate that you feel the need to do this. More power to you. However, flaming is unnecessary and poorly executed in this case.

Studies have shown that most people prefer the taste of AMD to the taste of Intel

Reply to shmeggegie

Quote :

You cant make an article stating once and for all why cl2 is better than 2.5 on ddr and use a system to run the test which for all intents and purposes ignores the effect of that cas latency benifit.

That would be like running a latency test on ddr ram using a 1 mb file on a processor with 2mb's of cache, [-peep-] POINTLESS.

I agree. I still maintain, however, that the performance of cl2 vs cl2.5 on a pentium 4 system is something that might interest pentium 4 customers, current or future. Therefore, I don't think the test should have been done on an Athlon so much as it should have been done on BOTH platforms. I understand that the Athlon would prove to be a more accurate testing platform if one were to write an article stating universally that one was better than the other architecture-wise. However, that would still leave questions for the Pentium 4 users who aren't necessarily sure how the technology would compare on their system. The most complete and accurate article would be the one that considers the most options, not just the standard or the p4.

Studies have shown that most people prefer the taste of AMD to the taste of Intel

Reply to shmeggegie

but its not a flamming post, just a basis thought. ll

if you know you don't know, the way could be more easy ...

Reply to labdog

Quote :

I agree. I still maintain, however, that the performance of cl2 vs cl2.5 on a pentium 4 system is something that might interest pentium 4 customers, current or future. Therefore, I don't think the test should have been done on an Athlon so much as it should have been done on BOTH platforms. I understand that the Athlon would prove to be a more accurate testing platform if one were to write an article stating universally that one was better than the other architecture-wise.

Fair enough, however I maintain that the article passed itself off as a test of ddr performance, not the p4's ddr utilization, and my analogy about the latency of ram on a 1 meg file using a processor with a 2mb cache is still valid.

You cannot run tests on hardwares archetecture when said hardware is being overridden by another piece of hardwares design, it just dosent make sense.

The test should have been done on both platforms, deffinatly, and also on non cpu controler simm testers even, or have a discussion of the differences between cl2 and 2.5 in a theoretical sense.

Of all possible routes vokkel took, he chose the worst!

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:

Reply to Matisaro

I completely agree.

Studies have shown that most people prefer the taste of AMD to the taste of Intel

Reply to shmeggegie

Forum Motherboards & Memory : Memory - Cl2 versus 2.5 official article discussion here.

Please mind

Focus On

Cl2 versus 2.5 official article discussion here.

Forum Motherboards & Memory : Memory - Cl2 versus 2.5 official article discussion here.

Please mind