good timings?

Ageeb · Sep 27, 2003

Are these decent timings?

1x256 Generic 2700 DDR
8,3,3,2.5 @ 210mhz @ stock voltage

I tried for x,2,2,2 and she didn't like it

I probably mistaken, but this seems nice for $30 ram.

NF7-S 2.0
2500+
Antec TC 550

I think that's all that's important...

g0dM@n · Sep 27, 2003

For 30 bucks, that's fine. If you want you can probably set it to AGRESSIVE timings and maybe up the voltage. you can push 2.5,3,3,6 maybe, but that ain't a BIG difference.

As long as you got less than CAS3, where you have 2.5... that is sufficient. On the other hand CAS2 owns. I have CAS2 on my Geil Golden Dragon dual-channel 210fsb (2x256)

Ageeb · Sep 27, 2003

Thanks for the quick reply. Maybe I'll monkey some more with it. I'm pretty sure it doesn't like CAS2.. boo. I'll hold for a bit and try proc now.

james.miller · Sep 27, 2003

you dont want 6,3,3.2.5. for the best performance, you should have 8,3,3,2.5. where the "8" is "3"+"3"+2. Mabey 7,3,3,2.5.

that first number should always be the second number+third number+2 (im sorry i cant remember the names right now lol). For exaple if you were running X,3,2,2.5, then "X" should be (3+2)+2 = 7. Having that number lower can cause serious memory error and possible hard drive corruption aswell.

as for CAS, generally the lower the better. 2.5 is good, 2.0 is better

I have CAS2 on my Geil Golden Dragon dual-channel 210fsb (2x256)

my twinmos pc3200 happily runs at 225mhz and over @ 8,3,3,2.0 DC

BTW welcome to the forums both of you

Ageeb · Sep 28, 2003

Very useful, thank you.

So far tonight I've found my mem is decent at 215 mhz, but cpu is little more stable at [email protected] @1.8v's or so. I have only stock cooling atm, and I can't do much more right now. As it is, my idle is around 45-48c. Another thing, voltage didn't matter for my mem. for the most part, anything above 215Mhz at any voltage caused instability (2.7,8, or 9). hmm...

I did try running 215x10.5 to no avail up to 1.9v cpu. Not sure the window for cpu voltage right now.

Anyhow this place rocks and thanks for the welcome

.

I'll try that little algorithm out you gave me.

LxOxSxI

Illah · Sep 28, 2003

james.miller said:
you dont want 6,3,3.2.5. for the best performance, you should have 8,3,3,2.5. where the "8" is "3"+"3"+2. Mabey 7,3,3,2.5.

that first number should always be the second number+third number+2 (im sorry i cant remember the names right now lol). For exaple if you were running X,3,2,2.5, then "X" should be (3+2)+2 = 7. Having that number lower can cause serious memory error and possible hard drive corruption aswell.

as for CAS, generally the lower the better. 2.5 is good, 2.0 is better

my twinmos pc3200 happily runs at 225mhz and over @ 8,3,3,2.0 DC

BTW welcome to the forums both of you

So basically, I'm at cas2-3-3-6 right now, I should switch to 3-3-8? For my future mem I was planning cas2-2-2-5 but by your logic 2-2-6 would be better off? Why is this?

--Illah

KidJethro · Sep 28, 2003

So basically, I'm at cas2-3-3-6 right now, I should switch to 3-3-8? For my future mem I was planning cas2-2-2-5 but by your logic 2-2-6 would be better off? Why is this?

I curious about the same question. My HyperX 3500 is at 2,2,2,5....2,2,2,6 would better?

Speed_Mechanic2 · Sep 28, 2003

Basically it depends. If you are running 2-2-2-5 or 2-3-3-6 and want to know if it will run better at 2-2-2-6 or 2-3-3-7 (2+3+"2"=7) then the only way you will find out for sure it to benchmark it. Open Windows using "msconfig" in Diagnostic Set Up, use SANDRA memory bandwidth benchmark, then reboot and set timings to 2-2-2-6 or 2-3-3-7 respectively and benchmark again. Then revert back to normal Windows set up or whatever.

Jognt · Sep 28, 2003

Speed_Mechanic2 said:
Basically it depends. If you are running 2-2-2-5 or 2-3-3-6 and want to know if it will run better at 2-2-2-6 or 2-3-3-7 (2+3+"2"=7) then the only way you will find out for sure it to benchmark it. Open Windows using "msconfig" in Diagnostic Set Up, use SANDRA memory bandwidth benchmark, then reboot and set timings to 2-2-2-6 or 2-3-3-7 respectively and benchmark again. Then revert back to normal Windows set up or whatever.

why 2+3+"2"=7? in james.millers explenation, the CAS latency isnt counted for, so for 2-3-3-7 it would be 3+3+2

and for 2.5-3-3-7 it would be 3+3+2, and for 3-3-3-7 it would also be 3+3+2 so why do you have 2 2's in your calculation?

Speed_Mechanic2 · Sep 28, 2003

RAS Precharge (tRAS) should be at least CAS Latency (CAS) plus RAS-to-CAS Delay (tRCD). A "buffer" of 1 or 2 should then be added, whether 1 or 2 is best depends on who your talking to. A CAS of 2.5 is rounded up to 3.0.

So with a CAS 2.0 and a tRCD of 3 then tRAS should be tried at 2+3+(0,1,2) 5,6,7 and benchmarked to find which one gives best bandwidth.

Jognt · Sep 28, 2003

hmm, your using the CAS in your calculation..

but if you ask me: tRCD + tRP + 2 = tRAS...
meaning: with a cas of 2 and a tRP of 3 and a tRCD of 3, it would be 3+3+2=8.....

g0dM@n · Sep 28, 2003

Alright, so you guys are saying that it is tRCD + tRP + (buffer) = tRAS, or tRCD + tRP + (CAS) = tRAS.

I'm assuming you guys mean the first one, where the buffer could be 0, 1, or 2?

Okay so let us take the RAM in my slower computer for example:
The manufacturer says that it runs 2.5-3-3-6, so why should I run it at 3 + 3 + (buffer of 2?) = 8?

You guys are saying that that's how I should run it, 2.5-3-3-8?

edit: P.S. Thanks for the welcome by the way, James-miller. I was part of overclocking.com forums as well, but I figured I'd give you guys a shot too. When I have the time I will post all the specs of both of my computers. You should be impressed by them, except for the 256 RAM in this one obviously, hehe.

Jognt · Sep 28, 2003

i do not know if you should set your memory to 2.5-3-3-8, since im only human (a) and i am also still learning

(b)

but if your manufacturer says it can handle 2.5-3-3-6 then by all means try it, but if it gives better performance with buffer

(i would not know which gives better performance.)

oh yes, i ment the first one

N803 · Sep 29, 2003

I can't speak to how precise this is, but Mushkin's site offers an explanation for this:

Memory, in many ways is like a book, you can only read after opening a book to a certain page and paragraph within that particular page. The RAS Pulse Width is the time until a page can be closed again. Therefore, just by definition, the minimum tRAS must be the RAS-to-CAS delay plus the read latency (CAS delay). That is fine for FPM and EDO memory with their single word data transfers. With SDRAM, memory controllers started to output a chain of four consecutive quadwords on every access. With DDR, that number has increased to eight quadwords that effectively are two consecutive bursts of four.

Now imagine someone closes the book you are reading from in the middle of a sentence. Right in your face! And does it over and again. This is what happens if tRAS is set too short. So here is the really simple calculation: The second burst of four has at least to be initiated and prefetched into the output buffers (like you get a glimpse at the headline in a book) before you can close the page without losing all information. That means that the minimum tRAS would be tRCD+CAS latency + 2 cycles (to output the first burst of four and make way for the second burst in the output buffers).

Any tRAS setting lower tRCD + CAS + 2 cycles will allow the memory controller to close the page “in your face!” over and again and that will cause a performance hit because of a truncated transfer that needs to be repeated. Along with those hassles comes the self-explanatory risk for data corruption. That one is not a real problem as long as the system is kept running but in case it is shut down and the memory content is written back to the hard disk drive, the consequences can be catastrophic. For the drive, that is.

g0dM@n · Sep 29, 2003

This site says that:

"Bottom line for DDR technology is that, as a rule of thumb, for best performance and stability, tRAS should never be less than the sum of tRCD + CAS Latency + 2 (or more) cycles."

...so:
tRCD + CAS + 2 cycles = rRAS

Now in my case I had 2.5-3-3-6 (manufacturer rating), which is actually wrong! This is seriously some stupid sh|t, come to think of it. Why the hell would they say that when you actually need an additional 2 clock cycle delay in order to transfer the eight quadwords, where the eight quadwords are effectively two consecutive bursts of four.

Therefore, I will change my RAS Pulse Width (tRAS) to:

3 + 2.5(round up to 3) + 2 = 3 + 3 + 2 = 8

where I used the equation above, and tRCD = 3 and CAS = 2.5 (but I assume you round this up to 3).

My timings are now currently 2.5-3-3-8

The funny thing is that this formula does not involve tRP (row access time) whatsoever. Wouldn't you think that it should involve that as well?

Keep in mind that timings are quoted (in order) as:
tCAC(CAS)-tRP-rRCD-tRAS. Notice that the formula disregards the second timing (tRP).

??????????

Illah · Sep 29, 2003

Gonna have to change that when I get home...

--Illah

shiltz · Sep 30, 2003

hmm, wonder if this could be the problem i'm having, I finshed my water cooled setup and went to O/C my CPU and I actualy lost 1000 points on 3d marks 2001 going from 3.0Ghz to 3.3Ghz, my timings are set at the manufactures spec which is 2-3-3-6, if i'm reading what was said in this thread right I should have it at 2-3-3-8, or try tightening it to 2-2-2-6 right?

Jognt · Sep 30, 2003

or use 2-3-3-7 since its CAS + tRCD + 2

so 2 + 3 + 2 = 7 ( dont know if its availeble though

g0dM@n · Sep 30, 2003

shiltz said:
hmm, wonder if this could be the problem i'm having, I finshed my water cooled setup and went to O/C my CPU and I actualy lost 1000 points on 3d marks 2001 going from 3.0Ghz to 3.3Ghz, my timings are set at the manufactures spec which is 2-3-3-6, if i'm reading what was said in this thread right I should have it at 2-3-3-8, or try tightening it to 2-2-2-6 right?

Yes. IMO, I would say 2-2-2-6 or 2-3-3-7. You will probably have to have the highest voltage setting to obtain 2-2-2-6. You might be better off just trying 2-3-3-7.

shiltz · Sep 30, 2003

higher voltage isn't a problem, that's the one good thing about the GeIL golden dragon, it's rated and warentied up to 3.1v, to bad my MB won't go that high, think the max it will go to is 2.8

good timings?

Member

Inactive Moderator

Member

Member

Member

Member

Member

Member

All round good guy

Member

All round good guy

Inactive Moderator

All round good guy

Registered

Inactive Moderator

Member

Member

All round good guy

Inactive Moderator

Member

Similar threads