How far did you guys push Q6600 (B3)
- Thread starter xx04201987xx
- Start date
Thor86
Diamond Member
- May 3, 2001
- 7,886
- 7
- 81
Overclocking is all luck of the draw, and varies with all cpus and varies with components used (ie: motherboard, psu, cooling).
B3 Q6600s run hotter than the newer G0 steppings, so I've only see these safely overclock up to 3.2Ghz on air with reasonable temperatures, and most should be able to get to 3Ghz on stock vcore voltages.
B3 Q6600s run hotter than the newer G0 steppings, so I've only see these safely overclock up to 3.2Ghz on air with reasonable temperatures, and most should be able to get to 3Ghz on stock vcore voltages.
3.33Ghz before the thermals were way out of control. Then I upgraded to water, dropped temperatures and clocked it higher. Running at 3.6 now and no problems.
PCTC2
Diamond Member
- Feb 18, 2007
- 3,892
- 33
- 91
Q6600 B3 @ 3.3GHz (366.67*9), 78C MAX, eVGA 680i-A1 motherboard and Ultra120 NOT eXtreme air cooling.
(NOTE: It is not in my sig b/c I have been having problems with the rig and it is now at stock. I think it's the 680i.... AGAIN. I wasn't able to get CPU-Z or screenies of 3.3GHz. Now it's time for a new water rig. )
BonzaiDuck
Lifer
- Jun 30, 2004
- 15,963
- 1,612
- 126
I'm still running my B3-stepping. AdairUSMC and I had some useful exchanges when he was running his.
The upshot of it is this. Through some lo-tech innovations, I was able to keep my upper-bound of TJunction core temperatures below 65C with room ambient at about 80F while running the processor at 3.20 Ghz and a VCORE setting of 1.41875V.
This was also pushing the memory a bit at 1:1 and very tight timings of 3,3,3,6,1T, and I had to volt the memory at 2.20V. They were Crucials, and they eventually gave out. I could've loosened the timings and dropped the voltage to 2.15. That's hindsight.
But I digress. To my point: with that sort of cooling regime, I might have pushed the VCORE a tad higher and obtained 3.3 Ghz using any multiplier at lower temperatures than others have reported. I just decided not to do it.
Since then, I've relegated my 3.2 Ghz profile to TXT-file notes, and run the system at either of these settings:
2.88 Ghz; 4:5 CPU : RAM; 4,3,4,9,1T with VDIMM 2.05V and DDR = 800 Mhz.
3.0 Ghz; 1:1 CPU : RAM; 3,3,3,8,1T VDIMM = 2.15V and DDR = 667 Mhz.
3.15 Ghz; 4:5 CPU : RAM; 4,4,4,10,1T VDIMM = 2.125V and DDR = 875 Mhz.
By far, the best memory/cache bandwidth benchmarks come from the 3.15 setting, with the 3 Ghz setting a close second. If I wanted to push the memory farther than 875 Mhz at 4:5 and boost VCORE and FSB accordingly for 3.3 Ghz, I suppose it would be fine. But I've noticed that using the same latency settings at 880 Mhz doesn't seem as good -- possibly an indication of a balancing-act between L2-cache and memory.
These 3.15 Ghz results are Everest Ultimate scores that were better than those at 3.2 and 1:1. The SiSoft Sandra scores, being buffered, are less impressive across the board. Even so, using Everest only as a basis of comparison between settings, the 9,750 MB/s "Read," 6,500 "Write" and 6,700 "Copy" scores blow away benchmarked DDR3 results in the Jan. 08 CPU-Power-User Magazine.
* [FOOTNOTE:] The burned-out Crucial Ballistix DDR2-1000's have been replaced by Crucial Tracer DDR2-800's, and hopefully I will RMA the Ballistix so that I can eventually try some new settings. Meanwhile, I await Penryn and even more favorable pricing for either the E6850 or the Q6700 (G0).
The upshot of it is this. Through some lo-tech innovations, I was able to keep my upper-bound of TJunction core temperatures below 65C with room ambient at about 80F while running the processor at 3.20 Ghz and a VCORE setting of 1.41875V.
This was also pushing the memory a bit at 1:1 and very tight timings of 3,3,3,6,1T, and I had to volt the memory at 2.20V. They were Crucials, and they eventually gave out. I could've loosened the timings and dropped the voltage to 2.15. That's hindsight.
But I digress. To my point: with that sort of cooling regime, I might have pushed the VCORE a tad higher and obtained 3.3 Ghz using any multiplier at lower temperatures than others have reported. I just decided not to do it.
Since then, I've relegated my 3.2 Ghz profile to TXT-file notes, and run the system at either of these settings:
2.88 Ghz; 4:5 CPU : RAM; 4,3,4,9,1T with VDIMM 2.05V and DDR = 800 Mhz.
3.0 Ghz; 1:1 CPU : RAM; 3,3,3,8,1T VDIMM = 2.15V and DDR = 667 Mhz.
3.15 Ghz; 4:5 CPU : RAM; 4,4,4,10,1T VDIMM = 2.125V and DDR = 875 Mhz.
By far, the best memory/cache bandwidth benchmarks come from the 3.15 setting, with the 3 Ghz setting a close second. If I wanted to push the memory farther than 875 Mhz at 4:5 and boost VCORE and FSB accordingly for 3.3 Ghz, I suppose it would be fine. But I've noticed that using the same latency settings at 880 Mhz doesn't seem as good -- possibly an indication of a balancing-act between L2-cache and memory.
These 3.15 Ghz results are Everest Ultimate scores that were better than those at 3.2 and 1:1. The SiSoft Sandra scores, being buffered, are less impressive across the board. Even so, using Everest only as a basis of comparison between settings, the 9,750 MB/s "Read," 6,500 "Write" and 6,700 "Copy" scores blow away benchmarked DDR3 results in the Jan. 08 CPU-Power-User Magazine.
* [FOOTNOTE:] The burned-out Crucial Ballistix DDR2-1000's have been replaced by Crucial Tracer DDR2-800's, and hopefully I will RMA the Ballistix so that I can eventually try some new settings. Meanwhile, I await Penryn and even more favorable pricing for either the E6850 or the Q6700 (G0).
Markfw
Moderator Emeritus, Elite Member
- May 16, 2002
- 26,437
- 15,563
- 136
Originally posted by: PCTC2
Q6600 B3 @ 3.3GHz (366.67*9), 78C MAX, eVGA 680i-A1 motherboard and Ultra120 NOT eXtreme air cooling.
(NOTE: It is not in my sig b/c I have been having problems with the rig and it is now at stock. I think it's the 680i.... AGAIN. I wasn't able to get CPU-Z or screenies of 3.3GHz. Now it's time for a new water rig. )
Well, mine runs 24/7/365 @ 100% on all 4 cores doing F@H, so I know its stable...
Edit 1.408 vcore.
I have mine up to 3.2GHz with an Ultra 120 Extreme and a 680i board. It can run a little too warm at times, though.
Mine is at 3300 on the ABIT IP35-E with Scythe Ninja Rev. B - idle at mid 30s and load in late 60Cs. Core volts at 1.38.
BonzaiDuck
Lifer
- Jun 30, 2004
- 15,963
- 1,612
- 126
Fact is -- serious reviews have shown B3's can be stable with air-cooling at 3.4 Ghz, but you will also push the voltage close to or beyond 1.45V.
The retail-box maximum for the Q6600 is 1.35V, and "default" voltages reported for it are around 1.372V. I'm not that eager to push my voltage too far beyond the "default," although it probably wouldn't do much harm -- in the "short-run" perspective of things.
I attended to the best air-cooling solutions, but I'd rather keep the cores at lower temperatures than what is still acceptable, and I'm not all that goosey about pushing the VCORE higher, either.
The retail-box maximum for the Q6600 is 1.35V, and "default" voltages reported for it are around 1.372V. I'm not that eager to push my voltage too far beyond the "default," although it probably wouldn't do much harm -- in the "short-run" perspective of things.
I attended to the best air-cooling solutions, but I'd rather keep the cores at lower temperatures than what is still acceptable, and I'm not all that goosey about pushing the VCORE higher, either.
I currently have my Q6600 at 3.6GHz (9x400) using 1.425 volts and it is Orthos stable (24 hours). I can get to 3.8GHz but it's not worth the extra voltage and heat.
Oops. Sorry, I didn't see that you wanted only B3 steppings. Mine is a G0. Sorry for posting here.
Oops. Sorry, I didn't see that you wanted only B3 steppings. Mine is a G0. Sorry for posting here.
Mine is at 3.1ghz, I'm using water cooling, it was hitting 85c on a Scythe Ninja which is way too hot for my taste, running at about 59c on water. Mine runs folding@home 24/7 as well.
My B3 is at 2.8 with stock voltage. It wasn't 100% stable at 3 GHz with stock voltage. Hovever, I had it up to 3.6 with added voltage.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 23K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 18K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Discussion Speculation: Zen 4 (EPYC 4 "Genoa", Ryzen 7000, etc.)- Started by Vattila
- Replies: 13K
Facebook
Twitter