RAM timings twk, will this cause sig heat

Have read OC sticky and would now like to pose some ques wrt to RAM timings..

I would like to tweak my RAM and try to optimise my memory but would like to ask advice from more experienced people firstly:

1. Will tweaking my RAM generate significant temps and contribute to overall sys temp, but especially CPU (if so I won't be doing it for now)?

The reason I ask this is I have a Prescott and was suffering CPU temp issues during summer (ambient ~31-32C), removed crusty PAD and refreshed TIM w AS5 (reduced 5C) but was still idling @55C and gaming (or load) would get me to 62-63C though as soon as stopped temps dropped very quickly to ~55C. Have seen this is about average for Prescotts from overclock.net and elsewhere and considering it being summer anxiety decreased. Heading to cooler months and (ambient 31C, low humidity) CPU@~49C, mobo 39C now, w cool nights CPU@46C.

My RAM is pict below and I have no idea of it's potential or lack off.

BIOS Options:

Advanced Chipset Features;

1. DRAM Clock/Timing Control [Press Enter]
2. AGP & P2P Bridge Control [Press Enter]
3. OnChip AGP Control [Press Enter]

Within each subsection (1):
1.
Current CPU Freq. [200Mhz]
Current DRAM Freq [200Mhz] all three are greyed out
Current DDR Freq [400Mhz]
Performance Mode [Disabled] enable to tweak RAM settings!
DRAM Freq [By SPD]
DRAM Timing Control [By SPD]
x DRAM CAS Latency 2.5T
x RAS Active Time (tRAS) 6T
x RAS Percharge Time (tRP) 3T
x RAS to CAS Delay (tRCD) 3T
MA 1T/2T Select [Auto]

So SPD (the Serial Presence Detect) is the currently selected option

Frequency/Voltage Control

4. Auto Detect DIMM/PCI Clk [Enabled]
5. Spread Spectrum [Disabled] manual says to disable if OC, so have for the moment
6. CPU Freq [Default] options are 100, 133 or 200 so can lower FSB freq but not raise!

These are my current settings and options. No options to tweak multiplier.
EVEREST Ultimate states:
CPU Multiplier min/max: 14/15
RAM timings [200Mhz 2.5-4-4-8 .(CL-RCD-RP-RAS)] or [166Mhz 2.5-4-4-7 (CL-RCD-RP-RAS)]

and if I do the memory Benchmark in EVEREST Ultimate is notes current RAM timings below.

I wont consider OC at the moment until thermal mangament is more under control and will use this link as guide for RAM timings:

 

Continued from above (locked out and Edit ran out as well)

RAM timings link:

http://www.tomshardware.com/2006/01/04/bios_from_a_to_z/

and read the relevant sects of the Holy Grail of RAM:

http://forums.pcper.com/showthread.php?t=320351

My current timings are below or is the other ways to ascertain this?

Will be running benchmark and stablity tests (BST) after each change and currently have memtest86, 3DMark03 and will source some others from the OC link but I suffer w dial up so big apps will take until 2007 to DL. Any good suggs here appreciated.

Timings Order: the link suggz to lower (tRCD) first then CL and finally (tRP) and that:

Eq.1 tRAS = CL + tRP + 2* where 2 is safety margin from tomsharware link, but Holy Grail differs and states:
Eq.2 tRAS = CL + tRCD + 2

Tons of info but I hope this will satisfy any obvious ques.

Main Ques:

Will there be heat issues?

Will my RAM do it okay?

What is the most appropiate order to approach this, e.g tRCD, CL and then tRP?

tRAS should be a product of?

What would be best for BST and how long?

How to ascertain actual current timings?

Any help greatly appreacited.

 

Well it looks like you have done your homework .....Iv been meaning to do a major update to the guide regarding temp increases when overclocking but I cant seem to make the time .. If your asking about ram temps dont worry about it at 3.0V and lower thare wont be any heat inless very poor air flow. As for your "Preshot" your temps sound very resonable to me...

 

Will there be heat issues? : Depending on your overvolt yes. If you dont overvolt no.

Will my RAM do it okay? : Again depinding on your ram overvolt. 3.0V and less is fine when running the ram "naked".

What is the most appropiate order to approach this, e.g tRCD, CL and then tRP? : The first timings make for the biggest boost in speed the last makes the least... Cas tRP tRCD tRAS in that order

tRAS should be a product of? : As for the holly grail vs. toms thing. tRP and tRCD are almost always the same.... In your case both are 4..

What would be best for BST and how long? : BST?

How to ascertain actual current timings? : CPUz should tell you.

 

just remember... there isn't much you'll gain in the "real world" from upping your timings. About the only thing you'll see get 'faster' will be a SisoftSANDRA bench, possibly 1-200 more 3dMarks, and if you do any Distributed Computing projects, you'll see about a 2-3% increase in productivity.... other than that, you won't see any benefits.

 

Winrar and Winzip (prolly any other compression utility for that matter) love fast/tight ram... But its up to you if you whant to shave half a secound off compressing an arcive . To me its not that that half of a secound realy matters. Then why do all the work to get it you ask? Because you can ..

As far as no "real world" benefits... From DDR 500 (3 4 4 13) VS DDR 500 (2 2 2 11) i get 20 FPS faster on average in any of my games....

 

Thanks ACE sorry if it was too long. BST is an acronym I dropped in for benchmarking and stability testing. I now have Prime95 (conceptually like it), memtest86, S&M and 3DMark03. From what I read more exp will run short intial test for first few tweaks then about 8 hrs to ensure stab.
Any suggz here? Which and how long?

NO OVERVOLTING, until I sqeeze another 5C permanently. Chuffed this morning, currently on 44C @ ambient 24C.

Okay have CPU-z but also EVEREST Ultimate 2006 below and will trust this just wasn't certain.

And yea 1,2, 5% at the moment if I can I will as long as it wont affect current thermal profile.

 

Sounds like you have it under control. I go for an hour on the stibility testers but thats me. Over night is best if you feel comfertable leaving it alone all night.. That is very cool for a "preshot" id go for more but its up to you..

 

Okay update on the methodology. I will start with the methods outlined in;
"Tight Timings vs High Clock Frequencies" @

http://www.tomshardware.com/2006/03/31/tight_timings_vs_high_clock_frequencies/print.html

i.e.
DirectX 9
3DMark01 SE Build 330
Resolution: 640x480, 32 Bit (to eliminate any graphics bottleneck)
Default Benchmark Run
Synthetic
SiSoftware Sandra 2005 Version 2005.10.10.69 SR3
Memory Bandwidth Benchmark
Everest Ultimate Edition 2006 Version 2.50.480
Memory Latency Benchmark
Applications
SuperPI Version: 1.1
8M

 

Damn hit Tab and it posted then cut me off and above post looks crap and unformatted!!!! Have done baseline SISandra 2005 complete benchmark report, EVEREST Ultimate 2006 memory cache benchmark, memtest86 for 1 run. Will aim for 2.0-3-3-7 from 2.5-4-4-7 and still considering the prolonged overnight stability test w all above conducted btwn each tweak and memtest86 for 1 run.

If interested here is an excellent link on benchmarking apps:

http://www.benchmarkhq.ru/be_cpu.html

Anyway will generate a basic before and after report, here at home, and if any interested will post it.

 

2.5-3-3-8 alright timings?
VALUE SELECT RAM (CROSIAR)

 

those timings look good. Same timings I'm using on my machine.

 

Well the timings look allright for an AMD I just want to squeeze a bit more out of my "Prescott" I noted you had slight OC in on that other threads as well, this is what I am avoiding, as previously stated had heat issues over summer and if you've exp thermal throttling kicking in and slowing sys performance down by about 10 years. Have to avoid heat until I can address this; new HS, mod front inlet, etc etc.

But "tightening" my timings can give me an perfomance boost (2-5%) wout heat issues, in comparison to "loosening timings" if you want to seriously OC but serious heat!!!!

Mine are at 2.5-4-4-7 but want to take to 2-3-3-7 (8?) and maybe lower will post EVEREST then for before and after.

 

No offence you can show me that, ACE has got his a lot lower and I'm sure you could. But what matters is how your system responds to benchmarking and stability testing. You also wanna decide want you want to do OC or Tighten.

 

i have oc'ed to 2.7 on stock cooler (245x11= 2.695 HTT@4x =980) and and stable with these timmings. i havent tried to tighten them up any.ill try though.

 

Have a look at memtest86 home page and use it that is were the first points of instability may become apparent, read it thouroughly but you may suffer the AMD test 5 or 8 issues. If you get heaps of errors in these tests it could be a false positive. You could also look at the link on benchmark apps and the one from tomshardware on tight timings vs clock freq.

 

Scott have a look at this:

http://www.playtool.com/pages/prime95/prime95.html

A good description to get out the whips and chains and make your memory system scream a bit.