Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Hardware

Memory Timings Analysis 159

Posted by CmdrTaco from the squeezing-out-the-fps dept.
keefe007 writes "It's generally known that smaller and more aggressive memory timings combined with higher clock speeds leads to higher performance, but for the most part, the increase in performance from tweaking each individual setting is relatively unknown. Perhaps in a bit too ambitious move, I set out to examine the impact of each individual memory timing and clock speed on overall performance. Find out the results of the tests at Techware Labs."
This discussion has been archived. No new comments can be posted.

Memory Timings Analysis More

Memory Timings Analysis

Comments Filter:

  • I've always found... (Score:4, Funny)

    by Lewie ( 3743 ) on Tuesday May 20, 2003 @10:04AM (#5998611)
    that the best time to install new memory is in between clock cycles.

  • Combinations, more! More!!!! (Score:5, Funny)

    by Whigh ( 663324 ) <afg205@yahoo.com> on Tuesday May 20, 2003 @10:05AM (#5998620) Homepage Journal
    What, only 289 combinations?!? I demand that all 4608 combinations be explored. Who knows what secrets of memory speed might be unleashed? Not that I'm willing to waste my time doing it.
    • What, only Windows XP was used for the tests as well? Why not run tests also against Linux and BSD on the same hardware too? Your result will vary...

      Testing just Windows is like test driving the Indy 500 ... driving a Pinto. You certainly won't be getting the full eXPerience.
      • "Testing just Windows is like test driving the Indy 500 ... driving a Pinto. You certainly won't be getting the full eXPerience."

        Funny you should mention that, I just played one of a pile of Indy 500 games on Windows XP. How ya doin on Linux over there?

      • "What, only Windows XP was used for the tests as well? Why not run tests also against Linux and BSD on the same hardware too? Your result will vary...

        Testing just Windows is like test driving the Indy 500 ... driving a Pinto. You certainly won't be getting the full eXPerience."

        Could be worse. He could have just tested Linux. How many extra FPS are you going to squeeze out of Tux Racer? Might as well test a speedboat in an olympic sized swimming pool. "I can drive nowhere, but I'm getting there real

    • I demand that all 4608 combinations be explored... Not that I'm willing to waste my time doing it.

      Hey, that's funny, but if a bunch of people did a handful of tests each, and someone was able to figure out the right statistical tools to compare those results, this might not be so far-fetched.

      • Only if there was A LOT of overlap. Different memory chips/main boards (even if they are the exact same model munber) would lead to slightly different results. If you were to set it up in such a way that there were many runs of the same test, and using a composit of ALL results you might get interesting results, but then your error bars would be larger and your results might be meaningless).

  • cas vs bus speed (Score:5, Insightful)

    by kochsr ( 144988 ) * on Tuesday May 20, 2003 @10:06AM (#5998629) Homepage
    so what i got out of that is that increasing the speed of the memory (from 133 -> 166) is a much larger difference than bumping down the cas latency. i think i'd rather have memory on a faster bus than at a lower cas then.

    but people will always say they have their stuff at the most agressive timings just to say that they are there, even though the average performance increase is only 0-2%
    • However, keep in mind that unless the bus speed of your CPU runs at the same 166 the memory does, you may even see a drop in performance!

      You often get better performance when the memory and cpu are running the same clock speed as then you don't end up with the CPU waiting on odd-cycles for the memory to pump data.

      You should run tests to verify, but hey, at least this is only two settings in stead of 289 or whatever.

    • Re:cas vs bus speed (Score:5, Insightful)

      by photon317 ( 208409 ) on Tuesday May 20, 2003 @10:41AM (#5998872)

      Yeah, that's a problem I have with this guy's test results. He's taken all the memory parameters which have very different purposes and effects, and rated them all on a scale of percentage increases in memory bandwidth.

      The important effect of dropping CAS latency is that it improves memory response time on a small request - it's not meant to really give a big boost to bulk bandwidth.

      You can think of the tradeoffs from spending a fixed dollar amount on clock speed boosts vs lowering CAS times as kinda like the difference between going RDRAM and going SDRAM. RDRAM had much higher bandwidth, but the latency sucked. SDRAM had lower bandwidth, but also had lower latency.

      So wrap it up - this test is uninteresting because it rated all those parameters based on how they affected bandwidth, when really only the clock speed speed has a significant impact on bandwidth - a lot of the other parameters are really more about latency and responding well to certain patterns of access.

      • Re:cas vs bus speed (Score:5, Informative)

        by barawn ( 25691 ) on Tuesday May 20, 2003 @11:57AM (#5999424) Homepage
        Definitely correct. Plus some of the other definitions were a little off (interleaving is essentially RAID for memory: it gets benefits because multiple devices can respond in parallel, rather than in series, so the latency penalty isn't incurred twice).

        What makes this terrible is the fact that there are latency measuring tools out there, lmbench specifically. It really wouldn't take that long to measure both latency and bandwidth.

        Considering the fact that this definitely would be interesting, it's a little annoying that he didn't do that.

        There are much more intensive memory benchmarks than Sandra. That's why it's a little annoying that Sandra's become so popular. There are other, easy to automate benchmarks that do a much better job. Sandra's useful, but not for this kind of thing.

        Just plain useless.
        • (interleaving is essentially RAID for memory: it gets benefits because multiple devices can respond in parallel, rather than in series, so the latency penalty isn't incurred twice).

          That's not entirely correct, either. He's talking about "bank interleaving" -- multiple banks on the same module all communicate through a single bus, so there's no benefit to having two banks read at the same time, as only one can return the data.

          To achieve "RAID for memory", in performance sense, you need channel interleavin
          • That's not entirely correct, either. He's talking about "bank interleaving" -- multiple banks on the same module all communicate through a single bus, so there's no benefit to having two banks read at the same time, as only one can return the data.

            Normal RAID works this way too. There's only one SCSI/IDE bus, so they can't read at the same time, because only one can return the data.

            Here's how you get a benefit:

            (Memory here has a latency of 2 clock cycles)

            1: Controller issues Read A to Chip A - Chip A
      • True. My framerate in QIII goes up 20-25% (depending on demo) on cas change from 333 to 222. All the other settings are the same meaning the actual bandwidth didn't change. That's on a dual PIII 1 GHz, VIA Apollo Pro chipset, PC133 SDRAM.
    • Lower CAS can give you a 5-10% performance boost, at least with PC133 ram.

      Tom's did a BIOS tweak article article about 2 years ago on this:
      http://www6.tomshardware.com/howto/20010725 /index. html

  • Results = Waste of Time? (Score:4, Insightful)

    by Baron_911 ( 664953 ) <{baron} {at} {insecure.net}> on Tuesday May 20, 2003 @10:09AM (#5998652) Homepage Journal
    Wow, considering all the settings that were tested, and the only improvment beyond 1% was the Clock Speed, seems like the rest of it was kinda a waste...
    • How can you call this a waste of time? Anything that let's you get 0.0001 extra frames per second in Quake 3 for free shouldn't be laughed at!
    • Wow, considering all the settings that were tested, and the only improvment beyond 1% was the Clock Speed, seems like the rest of it was kinda a waste...

      Well, was it more of a waste of time than playing some online game for 6 hours straight? Ever done that?

      While the outcome proves that there was no real increase, it was a worthy test. Would I have done it? No way. Was it a waste of time? Maybe, but was it a worthwhile waste of time, compared to other things? Probably. But the guy was right, until

    • Imagine this-

      A few ticks backwards of the clock
      One guy
      A limited number of brain cells
      One overgrown ego
      A couple of feathers
      Some glue
      Running at a good clip
      A decent sized cliff
      Some nice jagged rocks

      Everytime said guy runs off the cliff with a couple of feathers glued to his back he ends up a chunky wet spot on the jagged rocks.

      But that won't stop some other idiot from saying He just needed-

      More feathers
      Straps instead of glue
      To run faster
      A taller cliff
      Less pointy rocks

      Let's not get into the point of how it

    • Re:Results = Waste of Time? (Score:5, Informative)

      by CTho9305 ( 264265 ) on Tuesday May 20, 2003 @11:25AM (#5999204) Homepage
      I think he would have found more interesting results if he had chosen a different benchmark. The test he used only tested bandwidth, and latency was not a factor. However, most of the memory settings (other than clock speed) affect latency more than bandwidth. CAS is a major factor in latency. Had he used a benchmark that hit a few words at random memory locations rapidly, he would have seen the other effects of the settings he tweaked.

      If you've heard the quote, "Never underestimate the bandwidth of a 747 full of DVDs" (updated for modern times), you can see that it is an example of why bandwidth is not the only important factor. On the benchmark he used, a 747 full of DVDs probably would have scored pretty well.

      If you're going to play with latency settings, at least use a test that measures latency.
      • well, I think I'll stick with my actual RAM and settings because what I really don't want is a 747 full of DVDs to score onto my flat.

        Now, if you're talking about a flying saucer DVD, that's another think, but you've got to be pretty precise when it comes to direct landing in the slot-in dvd reader.

  • Full text of article in case of slashdotting (Score:4, Informative)

    by Anonymous Coward on Tuesday May 20, 2003 @10:11AM (#5998668)
    Memory Timings Analysis

    Review by Harry Lam on 05.16.03
    Test Ram provided by Crucial, MSRP: $26.00 (per stick)

    Introduction:

    The typical BIOS usually offers a varying number of settings directly related to memory: everything from timings to clock speeds. It's generally known that smaller and more aggressive timings combined with higher clock speeds leads to higher performance, but for the most part, the increase in performance from tweaking each individual setting is relatively unknown. Perhaps in a bit too ambitious move, I set out to examine the impact of each individual memory timing and clock speed on overall performance. The article that follows contains my experiences in this "memory benchmarking adventure" in conjunction with Crucial's PC2700 DDR RAM (and also gives a relatively good picture to the limits of this memory).

    I would recommend that anyone interested in learning more about memory timings take a look at this site. It gives a pretty good technical intro to memory timings.

    Testing/Methodology:

    Motherboard Selection:
    I decided to use the Soyo SY-P4X400 for testing, due to the flexibility of its BIOS in relation with memory timings, allowing me to change 10 different memory-related settings.

    Benchmark:
    To save on time and testing (all of the testing occurred over a 5 day period, with several hours of testing in each day), I picked only one benchmark: the memory test on SiSoft Sandra Professional 2003 v9.41 (SP1). I did notice that the initial few benchmarks on any configuration usually were significantly higher or lower than the "steady-state" score (the stabilized value that comes up after successive test runs of the benchmark in a row). To compensate for this, I selected the median score after the scores stabilized from successive benchmarks.

    Depth:
    I decided that 4,608 different combinations of memory timings on my particular test bench was a tad bit too much testing, and created a methodology which would get a look at the general increasing performance of memory timings but had the downside of having an uneven number of data points for timings that were deemed "less-significant (more on this later). VA Software is DEAD. This methodology simplified the number of combinations down to a mere 289 combinations (which actually still is extremely time consuming, considering that the test computer has to be reboot after testing each combination).

    I established Memory Speed (100, 133, 166), CAS Latency (3.0, 2.5, 2.0, 1.5), and Bank Interleave (Disabled, 2 Bank, 4 Bank) as the primary criterion for my benchmarking (as these usually are the settings that are most emphasized). The "less significant" memory timings (Trp, Tras, Trcd, DRAM Command Rate, DRAM Burst Length, Write Recovery Time, and DRAM Access Time) as a result received a less thorough testing.

    The general testing methodology is as follows:

    All combinations of Memory Speed, CAS Latency, and Bank Interleave were tested at the least aggressive memory timings, and once that was complete, I changed the first of the "less significant" memory timings to a more aggressive value (Trp was changed from 3T to 2T). I then repeated benchmarks for all possible combinations of CAS Latencies, and Bank Interleaves based on this new timing (12 total combinations). Slashdot really licks my nads. Once this was complete, I changed the value of the next "less significant" memory timing (Tras), and repeated another set of 12 combinations (keep in mind, I left Trp at 2T, the most "optimized" value). This process was repeated for each "less significant" memory timing, and then the entire set (of 96 different combinations) was repeated at an increased clock speed (for a total of 289 different combinations).

    As I stated earlier, this results in an uneven number of data points. For example Trp had 36 data points at 3T compared with 252 data points at 2T, and the reverse is true for DRAM Access time (252 to 36).

    Test RAM:

    Crucial was gracious enou

    • Re:nice troll. (Score:2, Funny)

      by Galahad ( 24997 )
      Yeah, people do read the posts. I found the following:

      • VA Software is DEAD.
      • Slashdot really licks my nads.

      Now if you really had a sense of humor, you would have changed the technical details. But then again, you're only an AC, right? Perhaps I'm expecting too much.

    • Indeed... Not even 50 posts and the site is slashdotted...

      Now I only wished you had posted the full text of the article, as you advertise, and not just the first page... And if you had to post only one page, the conclusions would have been more interesting. The above is pretty useless...

      If you are going to post the content of an article, please do it right :-)

      You get a good point, though, for posting as an AC and not being an karma whore :)

      • Re:Full text of article in case of slashdotting (Score:3, Informative)

        by Anonymous Coward
        Here is the results page, with a tiny bit of HTML formating restored for easier reading:
        -----

        Memory Timings Analysis

        Review by Harry Lam on 05.16.03
        Test Ram provided by Crucial, MSRP: $26.00 (per stick)

        Observations:
        • This can't be shown in the recorded results, but from my observations during testing, I noticed the general trend that as memory timings are set more and more aggressively, Sandra would reach it's steady-state score quicker and quicker (A gradual decrease from 5-7 trial runs to achieve a
    • since the _&%_&$#& /. filters do not allow a copy of page 2 i tried to make a mirror: (it is not /.éd now, it just loads slowly)

      <a href="users.domaindlx.com/leukhe/page2.htm">page2< /a>
      <a href="users.domaindlx.com/leukhe/page3.htm">page3< /a>

      By the way, the parent post is including items like "VA IS DEAD" in his text.
      • Don't forget to hold down SHIFT when you press your semi-colon key there buddy.

        That's what the preview button is for...
        • I first tried to copy and paste page 2 into a post. I thought it would look better in fixed font. The filters did not allow the table. I forgot to preview and now i have an ugly post that will be forever stored into the internet. But the mirror is there!

  • What beats me... (Score:5, Insightful)

    by jkrise ( 535370 ) on Tuesday May 20, 2003 @10:12AM (#5998673) Journal
    Is the fact that accountants and finance managers (decision makers in PC buying deals) talk as if they understand all these things better than sysadmins. SDRAM, DDRRAM, RambusRAM, L2 cache, on-chip cache and all that marketing crap is heavily used by these decision makers.

    Last year, I did a demo of a Via system with SDRAM and it did about 40% faster than a DDR-RAM board. The VP-Fin chap has become highly suspicious of any memory performance graphs or numbers, these days. And in true BOFH style, I've got decision-making rights on all PC purchases.

    Thanks to all the confusion.
    • I guess it depends on what sort of benchmark you were doing, but I find it very odd that an SDR system beat a DDR system unless something was done to stack the competition in favor of the SDRAM system. (Small loop that fits in CPU cache combined with a faster CPU for example.)

      For programs that fit in CPU cache, or spend a lot of time on the cache contents before moving on to a new dataset, DDR vs. SDR doesn't matter as much as programs that are not cache-efficient. (FYI, older FPS games, esp. Quake I/II

  • this guy (Score:5, Insightful)

    by daveatwork ( 655626 ) on Tuesday May 20, 2003 @10:13AM (#5998677)
    has was too much time on his hands. WAY too much time.

    Speaking as an engineer, I do hate buying new stuff because its cheaper to do so rather than spending time tweaking the old stuff, but 100's of combinations, for a few % increase? Even I would be perfectly happy in paying the money rather than loosing 0.05% of my life!

    • Re:this guy (Score:1, Funny)

      by Anonymous Coward
      This is like the mathematician that tried 69,376 combinations to find the most effective way to tie his shoes.

      At least this guy's actually done the research now. And it bolsters what is common wisdom.

    • Re:this guy (Score:3, Funny)

      by nai ( 465491 )
      This sounds like:
      - How much have you lost with this "Loose 20 pounds in 2 weeks" program?
      - Two weeks.
    • I think that you're overlooking exactly what work he put into it. By reading the analysis, you basically now know that:
      • Buy a system with the fastest bus speed
      • Don't worry about buying the fastest RAM based on CAS

      What does that mean? You can save $10-15 a stick of memory if you just buy CAS 2.5/3 instead of CAS 2 RAM (unless you really want that extra 2% increase of speed)

      I definitely think the write-up is on par with the earlier write-ups indicating that SDRAM kicked the snot out of RDRAM 3-4 years

      • 1: he didn't disable system cache, this could have cause the >theoretical performance increase.

        2: In his final analysis he failed to mention that there must be another bottleneck in the system causing the sub standard performance increases.

        3: He only tested memory transfer and not random access, page faults and all the other things that really slow your computer down.

        If your after max performance then your going to buy the best anyhow, if you not then a PC still using PC133 memory will be fine.

      • Re:this guy (Score:3, Informative)

        by Jackazz ( 572024 )
        ...is not smart.

        He only measured memory bandwidth...which does not exactly translate to real-world performance. He says there is no performance benefit from CAS2 as opposed to CAS2.5 or CAS3, but if you read Tom's Hardware [tomshardware.com] you'd know that CAS does have a drastic impact on overall performance. The benefit is just not in badwidth, it is in the time the processor has to wait from when it calls on the memory, to when it recieves the answer. The longer it is waiting for an answer, the longer it is sitting a

    • A few % increase can be the difference betwee 9 or 10 years, which is even more visible in parallel processing. If one machine is one whole minute faster a week, and you have 1000 machines. you just gained a few hours. Remember, it's all in perspective.

      As for a total waste of his time? Who cares? He was bored. We all have our hobbies. As long as he makes a living, and he comes out having had a good time, learning or what have you, be happy he did it. At least SOMEONE did.
      • If it is going to need to run for that long, you would be better off if you waited a couple of years and ran with a faster machine. Still, it is kind of interesting to see how things compare.
        • Uh.. no.. you can't predict that something will be fast enough to catch up on lost time and be fast enough to continue working.

          If something takes in the long run 10 years from now, or 9 years from 2 years from now.... well you do the math. But no one knows what to expect. So making the best out of what you have is always good.

    • I doubt he knew in advance there would only be a few % increase. What if he had found, for instance, that changing the DRAM Burst Length from 4 to 8 netted a 40% boost in throughput?

      His work is still valuable, if only to document (with numbers to back it up) that it's really not necessary to tweak these settings.

    • has was too much time on his hands. WAY too much time.

      According to you he has too much time on his hands. I myself enjoy doing this type of stuff, even if its only 0.01% increase its the satisfaction of achieving it rather than that extra 0.1 frame per second.

      But then again I am no engineer I just a pathetic physicist who likes to tinkle around with stuff I know nothing about.

  • Conclusion...UPDATED (Score:5, Funny)

    by fobbman ( 131816 ) on Tuesday May 20, 2003 @10:15AM (#5998697) Homepage
    Advanced memory timings, while beneficial in squeezing that last bit of performance out of your system, won't save your server from the Slashdot Effect.

  • Just bought some of that PC2700 myself (Score:3, Interesting)

    by graveyhead ( 210996 ) <fletch AT fletchtronics DOT net> on Tuesday May 20, 2003 @10:15AM (#5998700)
    I read several articles that said that PC3200 is not worth the price difference and that in many casses PC2700 is faster.

    Before I plunk down $$$ for PC3200, I wanted to know if it is worth it. I was hoping this article would help answer the question, but it looks like he is only testing various BIOS settings with a single DIMM, and not comparing memory with different access speeds. Any /.ers have experience with this? Is PC3200 worth the price?

    • Re:Just bought some of that PC2700 myself (Score:1, Informative)

      by Anonymous Coward
      PC3200 is faster if FSB can utilize it. PC3200 is recommended if you're using Athlon @ 400MHz FSB or Pentium4 @ 800MHz FSB.
      • not to mention with faster RAM that's guaranteed for faster everything then if you downclock you can bump the performance ever more. I'm using PC3500 though my CPU is only a 533 FSB (133x4) Bumping that to a 150Mhz (600 FSB) gets me an extra 300Mhz easily and my RAM is not nearly choked yet. Once I've built my water cooling kit I should be able to comfortably run the RAM at 166 (664 FSB?) effectively giving me a nifty 3Ghz CPU :) And my RAM still has room to play!!

    • Re:Just bought some of that PC2700 myself (Score:1, Interesting)

      by Anonymous Coward
      Read any of the motherboard shootouts on Toms Hardware, and you will find out that they did indeed find that in many cases PC2700 preformed the same as or better than PC3200 with exsisting chipsets (including the nForce2 and KT400A). I would hold off on the PC3200 until it becomes a definate performance leader.
    • It's only worth it if you're running a full 333MHz FSB. If you've got a processor in there running at 266MHz, and you're using a multiplier to have the memory bus be at 333MHz, you'll see almost no performance gain over PC2100.

  • I've found.. (Score:4, Funny)

    by grub ( 11606 ) <slashdot@grub.net> on Tuesday May 20, 2003 @10:26AM (#5998780) Homepage Journal

    I've found that dumping beer onto my computers' silicon memory has the same effect as dumping beer down my throat does on my carbon-based memory.

  • So, CAS has no effect on bandwidth???? (Score:2, Interesting)

    by Anonymous Coward
    Hmm, what a surprise.
    Sisoft Sandra memory bandwidth tests are good at exactly that, memory bandwidth tests.

    What would be much more interesting is how programs that rely on small memory latencies (especially scientific programs) depend on changing the CAS.

  • Other Benchmarks (Score:3, Insightful)

    by aliens ( 90441 ) on Tuesday May 20, 2003 @10:33AM (#5998829) Homepage Journal
    Synthetic Sandra are nice, but I'd rather he ran 3dmark2001. It's what I would've done ::)
    • 3D Mark will only provide a few (literally) additional points with different RAM timings. Overclocking the CPU or Video card will yield better results for gaming benchmarks. Application performance AND gaming will both be improved the better timings on the RAM though. Running my RAM at 2-2-2-5 gets me less than 50 more points on 3D-Mark. Nothing to write home about.
      • I'd like to have seen something like cachemem or sciencemark 2.0 membench latency numbers with the different timings.

        THAT would potentially have been an interesting article.
    • What he has proven is that under some conditions it might be possible to tweak some extra performance out of your memory AS LONG AS the only thing you do with your computer is run Synthetic Sandra.

      The first major problem I have is that he didn't test enough data points to prove that the differences in memory performance were actually due to the tweaks he did, and not thanks to some other environmental condition like sun spots, power surges, or what he ate for breakfast. His ENTIRE test suite only eaked out

  • Just for those who need more education. (Score:5, Informative)

    by SolidCore ( 250574 ) on Tuesday May 20, 2003 @10:36AM (#5998840) Homepage
    Memory Timing Explained. [ucr.edu]

    GamersEd.com [gamersed.com]

  • What do Compaq et al quote for bank interleave? (Score:3, Interesting)

    by swb ( 14022 ) on Tuesday May 20, 2003 @10:51AM (#5998937)
    The last server shindig I went to for Compaq had all these "advanced" memory options (hotswap, interleave, RAID for RAM [and for RAM only, it wasn't a solid-state disk system]).

    Does it really turn out that 4 way interleave is kind of bogus and only a 2-8% increase in performance? I suppose 8% might mean a lot, but on average it could be just 4% or so.

  • Wow, no shit (Score:2, Interesting)

    by stratjakt ( 596332 )
    Well, I'll get my flamebait mods, but what a no shit article.

    He finally concludes that memory clock speed has a significant effect on bandwidth, while CAS and other settings hardly have an effect at all. Something I've known intuitively all along, and anyone with a rudimentary understanding would know.

    In other words, when all those "super dooper case-moddin' overclockin' nothing-knowin' computer experts" payed an extra hundred bucks for stick of CAS 2 ram instead of CAS 2.5, they got ripped off. No surp
    • Actually, cas 2 is more than a few percent faster in applications that depend on latency. He ONLY used a bandwidth test - changing latencies OBVIOUSLY wouldn't have much effect there.

      Had he used a program that accessed many little bits of data throughout RAM (such as a scientific simulation) he might have found that CAS latency had a much larger role.
    • Well I like to buy the pretty Kingmax PC-150 sticks with the TinyBGA modules on them, but I only run them @ PC-100 or PC-133. I buy the best to reduce power consumption, heat and to increase reliability. I don't know many other geeks who have NEVER had to throw away a DIMM because it crapped out on them.

      I also underclock my athlon so it stays nice and cool, and run my hard drives in 'slow but quiet' mode.

      I don't know why people are so obsessed with tweaking every 1% out of their PCs, it's a good way to ki
      • Well, my last 2 cpus were a celeron 300a overclocked to 450, and a celeron 566 overclocked to 850. I can tell you, overclocking boosted the performance by far more than 1%. And I have not broken any parts.

        I can understand slowing things down to make them quieter, but if you think running your ram slower is saving significantly on repair costs you're nuts.

        • Well the celeron is a special case. It's a capable chip sold at a lower clock than it can run at. Also, you aren't overclocking your mainboard (fsb, ram, PCI, AGP), just bumping up the CPU multiplier. What you did was a 'clean' overclock. If you were running your FSB at 123Mhz and PCI at 36.5MHz you'd probably switch right back to the standard speeds, because your stability would plummet.

          As for my RAM, I like to know that I can move my memory chips to any machine without problems, and getting the Kingmax m
      • > Well I like to buy the pretty Kingmax PC-150 sticks with the TinyBGA modules on them, but I only run them @ PC-100 or PC-133.

        The other nice thing about buying overspecced components is that you have an upgrade path.

        I've got one stick of quality PC3200 RAM that I'm effectively underclocking, even when I overclock my cheapie P4 1.8A. With this CPU, 2700 would be just fine.

        But in a year or so, I'll get my money's worth when I hand the 1.8A to a friend and swap in a replacement CPU with a 166 or 2

      • Hey, do you live in a lead-lined house with an aluminum pyramid hat on all the time to block cosmic rays too? Your RAM, CPU, and hard drive will be worth pocket change in 2-3 years; who cares if by keeping the heat down you extended its life from 5 years to 10? Get real. If you bought 300$ computers every year you would have a faster machine on average than buying a 900$ computer every three years. Didja dink ah dat? Or do you just enjoy flushing money on expensive computer hardware so you can boast about f
        • No, I just don't like loud, angry sounding computers, and an athlon running full-tilt is just that. Occasionally I need to compile something HUGE (GCC, KDE, Mozilla, etc.) and I bring the clock up to full-speed. The other 95% of the time I'm more than happy running at 866Mhz.

          As for keeping the machine, I intend to give it to future offspring or a significant other down the line. My current girlfriend is happy with her Celeron 500, which is several years old. I figure by the time I move up to 'The Next Big
    • Your first statement is right on. CAS shouldn't affect _bandwidth_.

      Memory performance affects applications in basically 2 ways. Bandwidth and latency. He measured bandwidth. CAS changes latency.

      CAS 2 is better in many benchmarks than CAS 2.5, but almost certainly not in one that only measures bandwidth, since CAS primarily alters latency.

      Look at the real application benchmarks out there, rather than something synthetic, and you can see as much as 25% increase going from CAS 3 to CAS 2.

  • You mean...?!?!? (Score:5, Informative)

    by cbiffle ( 211614 ) on Tuesday May 20, 2003 @11:16AM (#5999136)

    Wow!!!!%#@

    You mean reducing RAM latency doesn't increase bandwidth?!?!#%!1 d00d!

    *sigh*

    This benchmark would have been vastly more informative if the guy had gotten his tests and statistics right. First, he needs to learn the difference between a median and a mean, which are very different. Second, actually testing latency might have been nice, considering that one of his independent variables is CAS latency. Not to mention the fact that the hardcoded pixel widths in the stats table are horribly wrong on a high-DPI system. People! The em is your friend!

    So basically what we have here is this:

    • Independent variables: bus speed (read: bandwidth), CAS latency, interleave (read: latency/bandwidth).
    • Dependent variables: bandwidth
    Quite frankly, if I had submitted this experimental design, my advisor would still have me tied to a table in the back end of the psych building. He's not measuring what he's manipulating, and throwing in a two-factor confound like bank interleave without compensating (though the article may be misleading) just skews the measurements.

    Ah, well. I'll go back to my completely untweaked Athlon and be happy. :-)

  • Testing method (Score:3, Insightful)

    by nuggz ( 69912 ) on Tuesday May 20, 2003 @11:20AM (#5999168) Homepage
    Someone needs to take a course, or read a book on Design of experiments.

    Then he could have gotten the same information with many fewer test runs.
    Also you could end up with interaction effects, which is nice. Maybe two settings have a greater or lesser effect.

  • I did my own tests... (Score:1, Interesting)

    by Anonymous Coward
    I did my own memory timing tests, I found that I got 40 more FPS in Q3 with DDR200@2-3-3-5 than at DDR166@2-3-3-5. When I lowered the timings to DDR166@2-2-2-5 I gained back all but a few of those 40FPS. To me that looks like memory timings have a significant effect in Q3 at least.
  • For some applications, like a main memory database, latency is key (bandwidth can be way below maximum for large number of small random read/writes). It looks like only large transfers were tested. None of his results suprise me. What I dont know is how certain settings affect latency.
  • I learned this the other day from an article at Tom's Hardware [tomshardware.com]. In retrospect it makes logical sense but I don't think it would have occured to me. We're sorta trained to think faster == more performance.
    Anyways, what the article discovers is that you'll get BEST performance when memory speed == FSB speed. In benchmarks they find that a Athlon 3000+ (333Mhz FSB) with DDR333 is faster than the 3000+ with DDR400 (or DDR444). So, mental note, when shopping for a system, don't bother paying extra for that fas
    • He concludes that running memory clocked at the memory speed = FSB is better than running mem speed > or < than FSB. He never says that running 400MHz DDR at 333MHz will be slower than DDR333 at 333. If you bother paying extra for that faster RAM, you can run it at 333MHz and it will be just as fast...

      The big thing about underclocking the bus speed, though, is that you can now overclock the latencies. You can make that CAS2.5 pc3200 a CAS2 pc2700 and tweak other latency settings, too. It also mea
  • Those of you out there running with >=1 gig of memory should be looking at ECC. How much this hits performance would be an interesting subject. I've heard ~5%
    At least I don't worry about cosmic rays; I use ECC!

  • Horrible test methods (Score:4, Informative)

    by boarder ( 41071 ) on Tuesday May 20, 2003 @12:50PM (#5999788) Homepage
    Seriously, this guy doesn't know how to run a good test setup.

    First off, he tested all these super specialized memory timings using a stick of RAM that was rated CL2.5 So he was overclocking it and stressing it when he ran a lot of the low latency settings tests. A better setup would've been to get the best darned stick of RAM and THEN test how the timings affect performance.

    Next, almost all of the timings he adjusted in the tests affect latency not bandwidth, but he used bandwidth as his ONLY benchmark. If a program is swapping small amounts of data, but VERY quickly and often, latency has more of an effect than bandwidth.

    Finally, he doesn't address asynchronous bus speed issues or how well some of his unattainable settings would work (because of my first complaint, his memory was unstable at the aggressive timings).

    I'm not a statistician, but it doesn't appear to me that he really understands some of the statistical methods for a good test. This is what I've garnered from reading other slashdot posts, at least.
  • Did he disable these caches? Seems to me if not then he really may not have tested anything. If he did then I can't see why there is so little difference between some of the tests.

    Of course, I've never run these benches so I don't know. I'm just asking.
  • It's generally known that smaller and more aggressive memory timings combined with higher clock speeds leads to higher performance

    Inventive use of the term "generally known"...
  • Apart from the lack of latency testing, why didn't he use programs to tweak this kind of stuff? There are programs that run from FreeDOS, that for example can adjust about all settings of your chipset.

    Then also use a good benchmark program that can run from FreeDOS and you're ready way faster. Not to bash any (more) real OS, but DOS boots way fast, it's easier to 'deploy' than making your own memory-benchmark-OS.
  • I'm fairly knowledgeable in computer architecture and PC hardware, but the one thing where I've always been lacking are memories.

    Could someone please post a link to a page with good overview of what PC2100/PC2700/... are, SDRAM/DDRAM/... On what kind of motherboards they should go ? What bus ? With which processors, etc...?

    And also what all the little tweaks in the BIOS are (latency, ECC, scrubbing...) I tried searching google, but it's always vendors' hype.

Slashdot Top Deals

If you aren't rich you should always look useful. -- Louis-Ferdinand Celine

Close