Found PC2-5300 RAM with CL=4. What is the catch?

I found G.SKILL PC2-5300 RAM with CL=4 (4-4-4-12 timing).
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231154

It is the only PC2-5300 RAM I could find that has 4-4-4-12 timing. The other PC2-5300 RAMs have 5-5-5-15. So, my question is: does it really run at 4-4-4-12 and what is the catch?

 

Do I need low-latency RAM?

It's sort of an old article (2004) but it defines the point on whether lower latencies are worth it or not.

 

Actually, my question is not about the cost. I just wonder if G.SKILL is hiding something. Normally, you have to trade between bandwith vs latency and now suddenly you can get both. What is the catch other than the cost? And how come we don't see reputable manufacturers like Corsair or Mushkin sell RAM with the same or better spec?

 

No catch. We've been waiting for this. There will be a small performance boost under those situations where getting data into / out of RAM is the bottleneck.

If you get this RAM then the Sandra memory benchmakr test result would be appreciated.

John

 

And we have been waiting for many people to post if 667Mhz at CL4 as far as I have seen no one has confirmed? This has been like 4 months so not like new? No one has done the slightest bit of effort to confirm and of course I know if running at 533Mz it will be CL4? I am troubled no one has posted, if the best thing since the "pooper scooper" would expect more posts? My argument is if it was people would have posted until then BS. Hi John, do you know what I mean?

 

Thanks, I will post the benchmark after I install it.

I found that Corsair has it too but it said that the RAM is optimized for MAC.
http://corsair.com/_productbriefs/Low_Latency_Mac_Memory_Performance.pdf
http://corsair.com/news/press_release.aspx?id=480559

Will I get the same benefit with PC (Sony FZ 4000 notebook)?

Corsair is $24 more expensive than G.SKILL. What do I get by buying Corsair? Which one should I buy?

 

Cheaper, you will have $24 less dollars.

 

Be careful. I started to buy this last year when it first poped up on newegg. I researched for a good while and found that a lot of laptop motherboard documentations say that it wont support 4-4-4 memory at ddr2667. Its usually very well hidden and hard to find.

 

That is true it will default to 5-5-5-15.

 

Well, u can change the timings by some tools e.g. Thaiphoon. I made mein to 667@4334 and havn´t met any problem.

 

OK, I just got one. I noticed that my system boots a bit quicker but that's probably due to upgrading from 2x1GB to 2x2GB rather than having RAM with shorter latency.

Here's the benchmark

Testbed:
Sony FZ4000 CTO
Intel Core 2 Duo T9500 2.6GHz
Intel GM965 chipset
2GB G.Skill PC2-5300 (4-4-4-12)
2GB Crucial PC2-4200 (4-4-4-12)
1GB Qimonda PC2-5300 (5-5-5-15)
Sandra Benchmark Free Version
Windows Vista Business SP1

4GB is detected by the BIOS.

CPU Timing according to CPUZ



Notice that only 1 Timing Table is available. This is a bit odd because usually I have more than one timing table for various memory speed. I don't know why I only have one. Does it mean that this particular G.Skill RAM is not backward compatible with slower system? It is also odd that Serial Number is not detected.

Sandra Benchmark
Memory Latency


Memory Bandwith


Cache and Memory


Please note that Sandra benchmark results could vary if you run it more than one time. So, I am not so sure about the accuracy of the benchmark. Moreover, I don't have 2GB PC2-5300 (5-5-5-15) stick. The benchmark could have been more fair if I could compare it directly to 2GB PC2-5300 (4-4-4-12) that I have. I also have a feeling that my G.Skill isn't as good as the other competitors. But this is the only PC2-5300 with 4-4-4-12 in the market. I think Corsair has one but it is too expensive and they said it is optimized for Mac.

Anyway, this is the benchmark I got. Any comments?

 

Thanks for this info. I'm a bit late noticing it. I'm also surprised to see that there's no SPD data for 266MHz. Sensibly this RAM would be CL=3 at 266MHz.

The Sandra benchmark results are very informative. They indicate that the lower latency doesn't substantially boost the RAM performance (100-150MB/s).

I'm dropping a link to this into the Dual Channel RAM Thread, so it's easier to find the results again.

John