stiull going on my ever long quest to understand OCing and the science behind it... as ive managed to figure out fsb and multipliers, and what they do relevant to the cpu and the rest of the system i thought id move on to ram and timings. partly because my ram is supposed to be good for timings and clocking with fsbs. now bear with me... sorry still a beginner... i thought it would help to look at my ram so that i can understand the timings involved. this is where ive found a problem....the holographic stickers on both my ram sticks (twin/dual set) state that these ram's are CMX512-3200C2PT. [/IMG] and the listed timings on their website for this set of sticks is 2.5-3-3-6 so i loaded up CPU-z and look at my current timings (untouched as i sure as hell dont know what to do yet). here i got confused as these timings dont match the ones supposedly for this ram on corsairs website. this is what CPU-z showed me [/IMG] now those aint right for these sticks right? so i had a look at what CPU-z said on the SPD chip settings. now heres the shock!!! [/IMG] anyone noticed or is it just me? the timings dont match this set of corsair ram........... ......... cuz its not CMX512-3200C2PT but CMX512-3200C2 so which one is it? [/IMG] or [/IMG] now is it just me being stupid or does the label on the tin not match the toy inside? and have i been done over by corsair in a good way? c2 is better than c2pt right? can anyone explain whats going on and have a stab in the dark what the actual default timings are for these puppies...... thanks everyone ps believe me its definetly helped me learn bout OCing this though if i am right... LOL
well those shots are awful quality and the difference between a 6 and an 8 in them is little. So you say corsair rate them for 2.5-3-3-6, and as i see it the cpu-z screenshot says 2.5-3-3-6 (or is it 2.5-3-3-8?). Many manufacturers relax the SPD timings so there is a good chance that the sticks will boot fine first time. Then you have to manually tighten them to the rated timings . my Ram is rated for 2.5-3-3-7 @ 200mhz iirc, but the SPD for them is 3-5-5-10 (crazy innit?), to make sure they will boot in a mobo that may not be able to handle the rated timings etc.
thanks a lot bud.... sos bout the pics (looked fine on my screen) but yes mine are set at 2.5-3-3-8. so you recon they can be tightened to 2.5-3-3-6 as stated on their website. and it still doesnt change the fact that they seem to be labeled differently?
righto. Im sure that they are what they are labeled, they just have a different SPD programmed into them then corsair rates them for . specs for CMX512-3200C2PT: Yes you could tighten them, but i recommend you use 2.5-3-3-7 (unless 2.5-3-3-6 gives better performance for some reason. you could try both and check with SiSoft sandra ) Tras (the last timing) should be set to CAS(the first.) + TRcd (the second) + 2 for best performance 2.5+3+2 = 7.5, but just use 7 since you can't set a half more info here .
Bear in mind also that with many chips the timings are a lot looser on AMD systems (there isusually a cryptic footnote to this effecty on their website) AFAIK the only Corsair that gives the same timings on Intel vs AMD is the XL series.
Yea, take little notice of SPDs (you will be as soon as you start clocking anyway as most likely youll want to loosen the timings to get higher FSBs then test the 2 settings to see which performs best). Its more about the manufacturer covering thier arse IMHO.
Yeah the SPDs are odd when I installed my 3200XLPT with 2-2-2-5 the timing were well off 3-6-6-8!! SO dont worry about it
well youve made me go and do it now.... its a slippy xpensive downhill slope now ...... ive started my first ickle overclocking, well tinkering with timers. but i need my timings checking cuz even though i noticed that when my timings were 2-3-3-6 it was absolutely flying along. i started on the default 2.5-3-3-8 and ran sisoftware sandra lite 05 bandwith benchmark. if it helps here are the results, cuz i can hardly decrypt them (figured most out) SiSoftware Sandra Benchmark Results RAM Bandwidth Int Buff'd aEMMX/aSSE : 3006 MB/s RAM Bandwidth Float Buff'd aEMMX/aSSE : 2803 MB/s Int Buff'd aEMMX/aSSE (Integer STREAM) Results Breakdown Assignment : 3028MB/s Scaling : 3024MB/s Addition : 2978MB/s Triad : 2994MB/s Data Item Size : 8 byte(s) Buffering Used : Yes Offset Displacement Used : Yes Bandwidth Efficiency : 94% (estimated) Float Buff'd aEMMX/aSSE (Float STREAM) Results Breakdown Assignment : 3022MB/s Scaling : 2835MB/s Addition : 2696MB/s Triad : 2660MB/s Data Item Size : 8 byte(s) Buffering Used : Yes Offset Displacement Used : Yes Bandwidth Efficiency : 88% (estimated) Performance Test Status Run ID : ADOLF on 02 March 2005 at 15:33:58 Memory Used by Test : 512MB NUMA Support : No SMP Test : No Total Test Threads : 1 SMT Test : No Dynamic MP/MT Load Balance : No Processor Affinity : P0 Page Size : 4kB Use Large Memory Pages : No Chipset 1 Model : ASUSTeK Computer Inc nForce2 AGP Controller Front Side Bus Speed : 2x 199MHz (398MHz data rate) Width : 64-bit Maximum Bus Bandwidth : 3184MB/s (estimated) Logical/Chipset 1 Memory Banks Bank 1 : 512MB DDR-SDRAM 2.5-3-3-8CL 2CMD Bank 2 : 512MB DDR-SDRAM 2.5-3-3-8CL 2CMD Channels : 2 Speed : 2x 199MHz (398MHz data rate) Width : 64-bit Maximum Memory Bus Bandwidth : 6368MB/s (estimated) Features (W)MMX Technology : Yes SSE Technology : Yes SSE2 Technology : No SSE3 Technology : No EMMX - Extended MMX Technology : Yes HTT - Hyper-Threading Technology : No Performance Tips Notice 5405 : System bandwidth appears FSB limited. Attempt to increase FSB. Notice 5008 : To change benchmarks, click Options. Notice 5004 : Synthetic benchmark. May not tally with 'real-life' performance. Notice 5006 : Only compare the results with ones obtained using the same version! Tip 2 : Double-click tip or press Enter while a tip is selected for more information about the tip. i then started lowring and playing with the timers.... something like: 2.5-3-3-8 2.5-3-3-7 (TMM recommended) 2.5-3-3-6 (TMM and corsair recommended) 2-3-3-6 and this was my final stopping point. unfortunately i did not take copy of the logs from sisoft but i did run the sisoft memory benchmark everytime, as i said i couldnt really make sense of it. well here are the reults from the 2-3-3-6 timings SiSoftware Sandra Benchmark Results RAM Bandwidth Int Buff'd aEMMX/aSSE : 3026 MB/s RAM Bandwidth Float Buff'd aEMMX/aSSE : 2835 MB/s Int Buff'd aEMMX/aSSE (Integer STREAM) Results Breakdown Assignment : 3042MB/s Scaling : 3039MB/s Addition : 3009MB/s Triad : 3016MB/s Data Item Size : 8 byte(s) Buffering Used : Yes Offset Displacement Used : Yes Bandwidth Efficiency : 95% (estimated) Float Buff'd aEMMX/aSSE (Float STREAM) Results Breakdown Assignment : 3037MB/s Scaling : 2869MB/s Addition : 2732MB/s Triad : 2705MB/s Data Item Size : 8 byte(s) Buffering Used : Yes Offset Displacement Used : Yes Bandwidth Efficiency : 89% (estimated) Performance Test Status Run ID : ADOLF on 02 March 2005 at 15:24:24 Memory Used by Test : 512MB NUMA Support : No SMP Test : No Total Test Threads : 1 SMT Test : No Dynamic MP/MT Load Balance : No Processor Affinity : P0 Page Size : 4kB Use Large Memory Pages : No Chipset 1 Model : ASUSTeK Computer Inc nForce2 AGP Controller Front Side Bus Speed : 2x 199MHz (398MHz data rate) Width : 64-bit Maximum Bus Bandwidth : 3184MB/s (estimated) Logical/Chipset 1 Memory Banks Bank 1 : 512MB DDR-SDRAM 2.0-3-3-6CL 2CMD Bank 2 : 512MB DDR-SDRAM 2.0-3-3-6CL 2CMD Channels : 2 Speed : 2x 199MHz (398MHz data rate) Width : 64-bit Maximum Memory Bus Bandwidth : 6368MB/s (estimated) Features (W)MMX Technology : Yes SSE Technology : Yes SSE2 Technology : No SSE3 Technology : No EMMX - Extended MMX Technology : Yes HTT - Hyper-Threading Technology : No Performance Tips Notice 5405 : System bandwidth appears FSB limited. Attempt to increase FSB. Notice 5008 : To change benchmarks, click Options. Notice 5004 : Synthetic benchmark. May not tally with 'real-life' performance. Notice 5006 : Only compare the results with ones obtained using the same version! Tip 2 : Double-click tip or press Enter while a tip is selected for more information about the tip. please could anyone tell me if those were right/good? or can someone make sense of the results and point me in the right direction..... thanks everyone
I tend to take little notice of benchmarks, especially sandra (unless it has changed recently) as its synthetic and holds little water in the real world. Try testing with superPi and prime95 and youll soon find an instability if there is one.
Looking good, you have increased your bandwidth. Its the 'Benchmark Results RAM Bandwidth Int Buff'd aEMMX/aSSE : 3006 MB/s RAM Bandwidth Float Buff'd aEMMX/aSSE : 2803 MB/s' bits that you're looking at. Getting close to my results However, just try setting it to 2-3-3-11 and see what it gives you. In most cases on nforce2 boards, 11 gives the best bandwidth; something to do with setting it to 11 changes a load of other settings that you cant. Also, as it's probably CH-5, you should be able to get it down to 2-3-2-x. If it doesnt boot, increase the VDIMM a notch. Winbond ICs love voltage...
this is what i have learned. The lettering PT after the C2 stands for the color of the heat spreader. PT stands for Platnum and no lettering stands for black heat spreader, The info came for Corsairs forums.
I have a stick of C2PT 1.2 and it has BH6 chips on it I think. Unfortunately, I probably shorted something when prying up the spreader as it no longer works . Wonder if I should try out their "lifetime" warranty?
As a note, the ram that you have is rated at 2.5 on AMD systems, which may be why it automatically defaults to that. It is rated as a CAS 2 on intel systems.
