Would it help to move from CL5 to CL4 DDR2?

[AstroManLuca]

I currently have an Athlon X2 with 4x1 GB of Corsair XMS2 DDR2-800, which is CL5. I'm planning on a mobo+CPU upgrade soon. When I do that, I could just move my current RAM over, but I'm thinking of instead getting new RAM. I figure I can use the opportunity to move to 2 GB modules, giving me future upgrade potential, and I can upgrade to CL4 while I'm at it.

Normally I'd just keep the RAM but the other aspect is that once I've done this upgrade, I'll have almost enough spare parts to assemble a whole new computer, which I can then sell. If I want to do that, I'll have to get RAM for it anyway, so I may as well use the upgrade as a second reason to get faster RAM.

So my questions are just as in the title. First, would it actually make any difference going from CL5 to CL4? Second, I've heard some people say that having four modules can actually cause more problems than two (I guess there's twice as much chance of one of them being bad?). Is there any good reason to stick to two RAM modules?

 

[taltamir]

less that 1% improvement in speed from going from CL5 to CL4.

going from 4 to 2 banks decreases: heat, power consumption, compatibility issues, overclocking issues (aka it gives better overclock).
It has no effect on speed though.

 

[BTRY B 529th FA BN]

Originally posted by: taltamir
less that 1% improvement in speed from going from CL5 to CL4.

going from 4 to 2 banks decreases: heat, power consumption, compatibility issues, overclocking issues (aka it gives better overclock).
It has no effect on speed though.

What about changing subtimings like with an Asus board & the Ai Clock Twister

 

[GarfieldtheCat]

Originally posted by: BTRY B 529th FA BN

Originally posted by: taltamir
less that 1% improvement in speed from going from CL5 to CL4.

going from 4 to 2 banks decreases: heat, power consumption, compatibility issues, overclocking issues (aka it gives better overclock).
It has no effect on speed though.

What about changing subtimings like with an Asus board & the Ai Clock Twister

Not worth it unless all you want to do is try and max out superpi or other benchmark.

 

[AstroManLuca]

Thanks for the help guys. I think what I'll do is, if/when I get around to selling the old machine and need to buy RAM for it, I'll just get slightly better RAM and put it in my new machine. Looks like there'd be no benefit to upgrading unless I have a buyer lined up for the old RAM.

 

[Blain]

With all the comments in these forums about how timings are meaningless these days,
I don't understand why this is still an issue.

 

[AstroGuardian]

Not much performance difference especially if your new rig is AMD which has lower RAM latency overall.

 

[BTRY B 529th FA BN]

Originally posted by: Blain
With all the comments in these forums about how timings are meaningless these days,
I don't understand why this is still an issue.

Yeah, that's the point. Everyone says its meaning-less but don't ask me, cause I won't tell you my experience playing UT3 before then after I changed Ai Clock Twister

 

[taltamir]

i dont say its meaningless... If I had a 1000$ phase change cooled CPU with triple SLI godbox that cost thousands of dollars, paying for higher end ram for a slight speed increase would be totally worth it. It is just something that is not very cost effective and thus not a useful way for an average person (aka, someone with a limited budget) to spend their money.

Iron man uses the fastest ram on the market though.

 

[AstroGuardian]

What kind of RAM does he use?

 

[TemjinGold]

Siege ram. :thumbsup:

 

[ghost recon88]

Originally posted by: taltamir
less that 1% improvement in speed from going from CL5 to CL4.

You have something to back that up? I noticed quite a difference changing my timings from 5-5-5-15 to 4-4-4-4.

 

[Blain]

Originally posted by: ghost recon88

Originally posted by: taltamir
less that 1% improvement in speed from going from CL5 to CL4.

I noticed quite a difference changing my timings from 5-5-5-15 to 4-4-4-4.

How do you know it wasn't a "Placebo effect"?

 

[ghost recon88]

Originally posted by: Blain

Originally posted by: ghost recon88

Originally posted by: taltamir
less that 1% improvement in speed from going from CL5 to CL4.

I noticed quite a difference changing my timings from 5-5-5-15 to 4-4-4-4.

How do you know it wasn't a "Placebo effect"?

Because of the fact that in doing so, it allowed me to get the WR for Q9550 in Super Pi 1M, wPrime 32M and 1024M, and PiFast

 

[Flipped Gazelle]

Originally posted by: ghost recon88

Originally posted by: Blain

Originally posted by: ghost recon88

Originally posted by: taltamir
less that 1% improvement in speed from going from CL5 to CL4.

I noticed quite a difference changing my timings from 5-5-5-15 to 4-4-4-4.

How do you know it wasn't a "Placebo effect"?

Because of the fact that in doing so, it allowed me to get the WR for Q9550 in Super Pi 1M, wPrime 32M and 1024M, and PiFast

Those are synthetic benchmarks. I'm sure taltamir was referring to "real world" performance - games, productivity apps and the like.

 

[BTRY B 529th FA BN]

Originally posted by: Flipped Gazelle

Originally posted by: ghost recon88

Originally posted by: Blain

Originally posted by: ghost recon88

Originally posted by: taltamir
less that 1% improvement in speed from going from CL5 to CL4.

I noticed quite a difference changing my timings from 5-5-5-15 to 4-4-4-4.

How do you know it wasn't a "Placebo effect"?

Because of the fact that in doing so, it allowed me to get the WR for Q9550 in Super Pi 1M, wPrime 32M and 1024M, and PiFast

Those are synthetic benchmarks. I'm sure taltamir was referring to "real world" performance - games, productivity apps and the like.

no matter how you cut it - it affects everything overall

 

[AstroGuardian]

Originally posted by: TemjinGold
Siege ram. :thumbsup:

I haven't laughed like this in hours hahaha

 

[taltamir]

every review I have ever read testing the real world performance difference of ram... from office applications, to file compression, to gaming performance...
This is what I base it on.

Also, siege ram FTW!