Memory question for A64 platforms

[BlindBartimaeus]

I am looking at building one...can't wait any longer...And since it has the onboard controller, it would see that the difference between the low latency and the higher 2.5 to 3 wouldn't have that big of an effect on performance. I have two 512 MB sticks of Kingmax 3500 and runs well at 2.5, and I was wondering if I keep them and use them in the new system or do I sell them and get lower latency stuff for the new computer?

Thanks

 

[myocardia]

Lower latency is always better, and always faster. Though, for instance, in gaming, you couldn't tell any difference whatsoever between 2-x-x-x and 2.5-x-x-x

 

[BlindBartimaeus]

I understand what you are saying but I am asking should I get some low latency stuff or is the difference negligible

 

[myocardia]

How am I supposed to know? You haven't told me anything yet but what processor you plan on buying. I told you that if you're only going to be using it for gaming, you'll notice no difference at all. Care to "open up" a little? What are your planned uses for this system?

 

[BlindBartimaeus]

I will use it for some light video editing and game playing. Have you built an A64 system?

 

[DAPUNISHER]

Originally posted by: BlindBartimaeus
I understand what you are saying but I am asking should I get some low latency stuff or is the difference negligible

The difference is indeed negligible, the best reason to buy tighter timing ram for A64 is to reach higher overclocks by relaxing them. If you don't plan to push it then stick with what you have. Here are a couple of links to benchmarks showing how little difference the timings make. I've seen a couple more articles that support this as well I'm just not certain where I saw them now.


Tom's
Ace's

 

[BlindBartimaeus]

Yeah...I am pretty sure they will do 230ish and if I get 230X11 I will be most pleased with an AN50R...that would put me at 2.5 gig on a 3000+ model.

 

[BlindBartimaeus]

Hey Punisher...I see you are at 2.35 gig with your A64. How do you like it? I am at 2.4 with an XP barton...did you notice the boost in performance over something like I have...same video card too.

 

[DAPUNISHER]

Right now I'm@2.4ghz 1.65v 5:6 LDT 2.5x and I have benched out both it and my Barton@2.43ghz and the Barton holds it's own very well in things such as ScienceMark, POVRAY, synthetic benchies, SETI, but for actual gaming the Barton gets beat down by a substantial margin.

 

[joe2004]

es and if you can get it to 2.6 GHz it will be as good as 64 at 2.4 GHz. It just doesn't make sense to buy AMD 64 before they come up with a discount or 939. Since you can work with Barton mobiles that have all the multipliers one would want.

 

[BlindBartimaeus]

Do others find this last statement to be true?

 

[Soulkeeper]

memory is too expensive these days, just keep your old memory for now

 

[BlindBartimaeus]

es and if you can get it to 2.6 GHz it will be as good as 64 at 2.4 GHz. It just doesn't make sense to buy AMD 64 before they come up with a discount or 939. Since you can work with Barton mobiles that have all the multipliers one would want

I meant that comment...sorry for the confusion

 

[cowdog]

On my A64 setup, I have tried both BH-5 based Mushkin memory at 6-2-2-2 and some much lower priced Corsair XMS PC-3200 at 6-3-3-2. I didn't really notice a difference, although I didn't run any kind of rigorous quantitative benchmarks. The biggest difference I notice is that I need to back off the timings on the non-BH-5 memory for overclocking, while I can overlcock the BH-5 memory to the same levels at 1:1 with spd timings. The difference in performance there seems minimal. For example, the one measure I do have is SuperPi. For 1M, I get 37sec with the Mushkin at 6-2-2-2 and 39 sec with the Corsair at 8-4-4-2 (I think). Aside from that, I really can't tell the difference.

I have 2 sets of coveted BH-5 memory (2x512 Corsair LL and 2x256 Mushkin Lv 2), and although I am very fond of that memory, I can't imagine buying those or related "high-end" memory sticks at current prices.

BTW, for my real-world computing and gaming, the stock or overclocked A64 is hands down superior to my overclocked Barton.

 

[BlindBartimaeus]

So you would say it is noticeable...perhaps I won't build a mobile athlon major overclocker

 

[DAPUNISHER]

Originally posted by: BlindBartimaeus

es and if you can get it to 2.6 GHz it will be as good as 64 at 2.4 GHz. It just doesn't make sense to buy AMD 64 before they come up with a discount or 939. Since you can work with Barton mobiles that have all the multipliers one would want

I meant that comment...sorry for the confusion

For everything I've tried besides gaming I'd agree with Joe, but for gaming the Barton even@2.6ghz is not going to touch a A64@2.4ghz.

 

[cowdog]

I have a boring "regular" Barton 2500+ where my max stable 24/7 overclock is approx. 2350MHz. If one can reliably get 2600MHz with a mobile Barton, then I'm sure that would close the gap.

I get temped by the mobile Bartons, but I already have one too many nForce2 motherboards and two too many XP 2100+ cpus. If I picked up a mobile Barton, then I'd end up with one too many Bartons... I have extra memory, chipset coolers, and so on. What on earth do all you regular upgraders do with all your "old" equip? I only have use for so many computers, and I seem to like to doink around with the shiny new stuff as it comes out.

Anyway, what I like about the A64 is that it runs super cool and, as has been mentioned many times, is great for gaming. I also do a lot of quantitive/science/geeky computing, and for those applications I really like the performance of the A64 in both 32-bit Windows and 64-bit Linux. For example, I have some metacode that runs on a statistics package (R) that takes over 15 minutes per run on my Barton and about 1 3/4 minutes on my A64. I run apps like that many times each day, and the time savings is significant. I have plans for more whacky, monster data blasting apps, but I will likely move on to a multi-processor workstation (that will be a "work" computer unless I fail to acquire the needed funds).