If I can run DDR 400 at 150 fsb@3:4 or 160@4:5 with my 845G board, which of the two modes should I be using? Will overclocking my cpu from 150 to 160 but same DDR just going to cost me stability and shorter system lifetime without any gain in overall performance?
Poll: Which ratio is best at same DDR? 3:4 or 4:5?
- Thread starter NanoMem
- Start date
If you can do it 160@4:5 should be better, but so far I haven't been able to get my 845G board (Epox 4G4A+ with 2.26B CPU) to run above 156 FSB. Still working on it though. Maybe it'll do it with a 1.6A, which I'll try later today.
I realize fsb@160 with 4:5 is the obvious choice in terms of fsb only but sandra mem scores suggest fsb@150 with 3:4 are slightly preferred by P4/845G. Please note that you can test the two ratios under other DDR's than just 400.
If you're using 3:4 now, take your FSB divide it by 2.5 and multiply the result by 2.66 to get your FSB for 4:5 mode. Example: If fsb=133, ratio=3:4, DDR=2*[133*(4/3)]=354, the above calculation of FSB for 4:5 would be (133/2.5)*2.66~141. Verification, at FSB=141, ratio=4:5, DDR=2*[141*(5/4)]~353. Both sandra scores would about 8*354 or around 2800 in this case.
Conversely, if 4:5 is in use, take the FSB divide it by 2.66 first and multiply result by 2.5 next to obtain equivalent fsb for 3:4 mode.
Example: If FSB=160, ratio=5:4, DDR=2*[160*(5/4)]=400, the new FSB for 3:4 mode is (160/2.66)*2.5~150. Verification, at fsb=150, ratio=3:4, DDR=2*[150(4/3)]=400. In this case, sandra should read about 8*400 or 3200 for both modes.
This exercise should shed some light on two issues. Is the gain of few extra FSB's in going from 3:4 to 4:5 for the same DDR really worth it? Would the presence of 4:5 ratio in BIOS be important to 845D/845E boards if 845G results are extrapolated?
If you're using 3:4 now, take your FSB divide it by 2.5 and multiply the result by 2.66 to get your FSB for 4:5 mode. Example: If fsb=133, ratio=3:4, DDR=2*[133*(4/3)]=354, the above calculation of FSB for 4:5 would be (133/2.5)*2.66~141. Verification, at FSB=141, ratio=4:5, DDR=2*[141*(5/4)]~353. Both sandra scores would about 8*354 or around 2800 in this case.
Conversely, if 4:5 is in use, take the FSB divide it by 2.66 first and multiply result by 2.5 next to obtain equivalent fsb for 3:4 mode.
Example: If FSB=160, ratio=5:4, DDR=2*[160*(5/4)]=400, the new FSB for 3:4 mode is (160/2.66)*2.5~150. Verification, at fsb=150, ratio=3:4, DDR=2*[150(4/3)]=400. In this case, sandra should read about 8*400 or 3200 for both modes.
This exercise should shed some light on two issues. Is the gain of few extra FSB's in going from 3:4 to 4:5 for the same DDR really worth it? Would the presence of 4:5 ratio in BIOS be important to 845D/845E boards if 845G results are extrapolated?
man is that a confusing explaination!
4:5 = +25%
3:4 = +33%
4:6 = +50%
3:5 = +66%
3:6 = +100%
isnt that easier?
4:5 = +25%
3:4 = +33%
4:6 = +50%
3:5 = +66%
3:6 = +100%
isnt that easier?
4:5 should always be better because your processor will have to be running faster besides running a higher FSB.
I apologize for not being clear. My intention was not to compare any raw increase of FSB's due to various ratios, but to only check the effect of that very specific condition of an "U N C H A N G E D" DDR while switching from 3:4 to 4:5 and only gaining 6.4% FSB's in the process. If I may stress again, this test requires that very special circumstance of an "I D E N T I C A L" DDR for both ratios.
Although I can explain the exact source for 6.4%, I recommend to verify the following. To go from 3:4 to 4:5 and keep same DDR of 354, FSB needs to be increased from 133 to 141. Similarly, if you want to run DDR400 switching from 3:4 to 4:5 requires FSB to be changed from 150 to 160. The same 6.4% increase will always apply to FSB if you want to switch 3:4 to 4:5 and keep DDR "U N C H A N G E D".
Is that 6.4% increase in FSB worth risking stability if I end up at the S A M E DDR just by switching from 3:4 to 4:5???
Although I can explain the exact source for 6.4%, I recommend to verify the following. To go from 3:4 to 4:5 and keep same DDR of 354, FSB needs to be increased from 133 to 141. Similarly, if you want to run DDR400 switching from 3:4 to 4:5 requires FSB to be changed from 150 to 160. The same 6.4% increase will always apply to FSB if you want to switch 3:4 to 4:5 and keep DDR "U N C H A N G E D".
Is that 6.4% increase in FSB worth risking stability if I end up at the S A M E DDR just by switching from 3:4 to 4:5???
I was just going to post a question about what does it all mean,thank's very much THUGSROOK.
No wonder when I set it to 4:6 at 133mhz it didnt finish booting.
DDR chart . . . 100fsb . . . 133fsb . . . 150fsb . . . 166fsb
4:5 = +25% . . . 250ddr . . . 333ddr . . . 375ddr . . . 415ddr
3:4 = +33% . . . 266ddr . . . 354ddr . . . 399ddr . . . 442ddr
4:6 = +50% . . . 300ddr . . . 399ddr . . . 450ddr
3:5 = +66% . . . 332ddr . . . 442ddr
3:6 = +100% . . 400ddr
4:5 = +25% . . . 250ddr . . . 333ddr . . . 375ddr . . . 415ddr
3:4 = +33% . . . 266ddr . . . 354ddr . . . 399ddr . . . 442ddr
4:6 = +50% . . . 300ddr . . . 399ddr . . . 450ddr
3:5 = +66% . . . 332ddr . . . 442ddr
3:6 = +100% . . 400ddr
Thugs, I guess your approach is not bad except I would've transposed the matrix and only focused on our two ratios of interest.
FSB DDR@3:4 DDR@4:5
============================
100 267 250
106 283 265
113 301 283
120 320 300
128 341 320
136 363 340
145 387 363
154 411 385
164 437 410
174 464 435
185 493 463
197 525 493
***************************
133 355 333
150 400 375
160 427 400
166 443 415
Consecutive FSB values are derived according to what we call a "geometric sequence" where a0=100 and r=1+1/16 or 1.0625. Notice how each DDR value at 3:4 is equal to the value of the next row in the other column under 4:5 except of course for special FSB values of 133, 150, and 166 added at the end. Pick a DDR from the second column, read the FSB @ 3:4. To find corresponding FSB for 4:5, locate same or closest DDR value in the third column and read off its FSB. As you can see the range is not small, and you can perform this test at all kinds of DDR's.
So far, my own findings reveal that once my maximum DDR is hit while fishing for it with 4:5, I am actually better off relaxing my FSB to the value @ 3:4 for the same maximum DDR. 4:5 looses its importance once RAM maxes out (is oc'ed to the max) but 4:5 served me as a probing tool for locating that evasive max DDR. DDR RAM overclocking is the last breath that can be squeezed out of the P4/845 combo and that is probably the main reason behind the bonus of 4:5 in the BIOS of 845G imo. The same performance can be achieved with 845 boards without 4:5 as long as the maximum DDR is reproduced with 3:4. With 1:1 only, your cpu becomes the culprit preventing you from hitting max DDR cuz just to hit DDR400 you'd need to drive your cpu all the way to fsb 200! 3:4 is just not any other ratio but is the key for simultaneous FSB and DDR overclocking imo.
FSB DDR@3:4 DDR@4:5
============================
100 267 250
106 283 265
113 301 283
120 320 300
128 341 320
136 363 340
145 387 363
154 411 385
164 437 410
174 464 435
185 493 463
197 525 493
***************************
133 355 333
150 400 375
160 427 400
166 443 415
Consecutive FSB values are derived according to what we call a "geometric sequence" where a0=100 and r=1+1/16 or 1.0625. Notice how each DDR value at 3:4 is equal to the value of the next row in the other column under 4:5 except of course for special FSB values of 133, 150, and 166 added at the end. Pick a DDR from the second column, read the FSB @ 3:4. To find corresponding FSB for 4:5, locate same or closest DDR value in the third column and read off its FSB. As you can see the range is not small, and you can perform this test at all kinds of DDR's.
So far, my own findings reveal that once my maximum DDR is hit while fishing for it with 4:5, I am actually better off relaxing my FSB to the value @ 3:4 for the same maximum DDR. 4:5 looses its importance once RAM maxes out (is oc'ed to the max) but 4:5 served me as a probing tool for locating that evasive max DDR. DDR RAM overclocking is the last breath that can be squeezed out of the P4/845 combo and that is probably the main reason behind the bonus of 4:5 in the BIOS of 845G imo. The same performance can be achieved with 845 boards without 4:5 as long as the maximum DDR is reproduced with 3:4. With 1:1 only, your cpu becomes the culprit preventing you from hitting max DDR cuz just to hit DDR400 you'd need to drive your cpu all the way to fsb 200! 3:4 is just not any other ratio but is the key for simultaneous FSB and DDR overclocking imo.
dont forget that once 166fsb becomes official it will need 5:6, 5:7, and 5:8 ratios for 400, 433 and 466DDR.
but chances are that DDRII will be out by then. (DDRII only runs on a 1:1 ratio so far - 333 and 400DDR)
but chances are that DDRII will be out by then. (DDRII only runs on a 1:1 ratio so far - 333 and 400DDR)
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter