lopri
Elite Member
Overall I feel it's easier to OC on Intel platform than on AMD platform. Or better put, it's simpler. In A64 OC'ing, as well as finding the limits on your CPU, memory, and chipset, respectively, and you also had to find the sweet spot with balance. It often took days to find that sweet spot with countless trial-and-error's. On the other hand, with Conroe on 975X it took me less than 24 hours to find the balance.. (well, actually imbalance might be a better description in this case) Being easy should be a good thing and it indeed is, but I should admit it's not as fun as OC'ing Opterons. Hopefully it won't take too long for motherboard manufacturers to give us some seriously capable boards so that we can truly push these excellent CPUs.
http://img76.imageshack.us/my.php?image=e6300retailfc5.png
Currently installing Windows.. 😀
--------------------------------------------------------------------
Edit: OC'ing in progress - Will update as it improves.
OK so I started having fun. Well, not so at the beginning. My setup is:
Intel Core 2 Duo E6300
P5W-DH Deluxe
Xtreme DDR2-667 1G x 2 (good up to DDR2-1080 / 4-4-3)
GeForce 6200TC
CNPS 9500
I booted the system without touching the BIOS other than disabling onboard sh*ts, etc., and raised the clocks little by little. (more on this later) So settings usually related to overclocking are at 'stock' configuration. (VCore, VDIMM, etc.) Every 50~100MHz I run either dual SP 32M or 5 mins of dual Prime95, then go up another notch. Findings:
1. Temps are extremely low @stock VCore. On this board, Pentium D 930 @4GHz used to hit 70C easily on load, and I've personally verified it with a thermo-gun. 😀 My E6300's idle temp basically equals to ambient temperature. (24C) This is a never-before-seen phenomenon to me.
2. I've stopped raising clocks @2.4GHz for now and currently running dual-Prime95 for about 40 mins. CPU Temp = 43C
http://img157.imageshack.us/my.php?image=2400apt0.png
http://img157.imageshack.us/my.php?image=2400btd8.png
3. The major obstacle so far has been the ASUS board. (What a surprise) At first I thought ASUS finally gave us a DFI-like board, but boy, was I wrong. You can still expect the usual ASUS behaviors from this board. Sure it's much improved (shouldn't have?) but expect to go through voltage fluctuation, warm-boot, cold-boot, having to raise FSB gradually, doing things on its own, etc. Again, it's much, much improved upon, say, A8N32-SLI, but it's still there and it's simply annoying.
4. At least ASUS implemented great BIOS recovery. I haven't had to clear CMOS yet.
Does anyone know what "FSB Termination Voltage" is and what it does, in the Award BIOS? Please post or PM me. I'd much appreciate it.
-------------------------------------
Edit 2: Super PI 32M @2.5GHz / Stock VCore
http://img195.imageshack.us/my.php?image=2500rf1.png
------------------------------------
Edit 3: Dual-Prime95 currently running @2.8GHz (7x400), STILL STOCK VCORE
http://img97.imageshack.us/my.php?image=2800aft0.png
-----------------------------------
Edit 4: Going to bed leaving the system running 2 instances of Prime95 @2.94GHz. Will report back tomorrow.
Edit 5: I'd consider the following stable.
http://img225.imageshack.us/my.php?image=2940prime4hrs3ed1.png
Also I could raise the FSB up to 430, but that required running memory with speed less than 1:1, which should be avoided on Intel platform. When I have more time, I will run the Prime95 again just to make sure.
Basically it's my first experience of motherboard (FSB, to be exact) limiting a CPU OC'ing. For the above Large FFT run, I had given 2.25VDIMM for memory. Had I given 2.30V, it'd have run much longer, is my guess. What happens is the northbridge is stressed by high FSB and memory (frequency/timing both), so even though CPU and memory have more potential, I won't see it on this board. Not that this board is bad in any stretch. This is the best ASUS board I've ever purchased, by far. I will make it short and easy:
1. E6300 is just like Socket 939 Opterons. 50% OC'ing seems easy.
2. Your motherboard's capability of high FSB will be the most limiting factor in OC'ing this low-multiplier CPU.
3. MCH (northbridge) gets stressed by both high FSB and memory.
4. This results in another problem - What if you won't be able to fully utilize your expensive memory because the memory controller on MCH can't handle it? This is a similar situation as in A64 OC'ing, where often times IMC used to be limiting memory performance. My sticks could do DDR2-800 @2.2V and DDR2-1080 @2.3 when FSB was under 300.
http://img95.imageshack.us/my.php?image=ddr21080ig9.png
But with the kind of FSB required for E6300 OC'ing, the story changes. I needed 2.3V for DDR2-840, and also had to loosen the timings. With anything higher, system just wouldn't boot.
5. On a side note, I'd strongly suggest an aftermarket cooler. It could be an isolated case but the IHS on my E6300 was very rough and uneven.
6. E6300 runs very very cool. Granted my system is an open-air system, the load temp was around 50C during dual-priming @2.94GHz.
7. P5W-DH users, make sure to disable "Hyperpath 3" when overclocking. I learned it in a hard way.
Now it's a benching time!
Small update: I'm currently testing @3010MHz (7x430) with dual-Prime95 Small FFT. While I'm happy the board is capable of doing 430HTT, it doesn't really mean much performance wise since I had to lower the memory speed below the FSB. (4:3) Will link to a pic after the Prime continues for at least a couple hours. VCore = 1.35V
-----------------------------------
OK here you go.
http://img220.imageshack.us/my.php?image=3010smallfft1nn5.png
http://img220.imageshack.us/my.php?image=3010smallfft2fr8.png
-----------------------------------
LOL once I gave up 1:1 the FSB goes up higher.. Booted @7x440. FSB:Memory=4:3
http://i35.photobucket.com/albums/d177/PenguinBelly/3080SmallFFTstart.png
SOMEONE PLEASE LINK ME TO A VMCH MOD TUTORIAL!
http://img76.imageshack.us/my.php?image=e6300retailfc5.png
Currently installing Windows.. 😀
--------------------------------------------------------------------
Edit: OC'ing in progress - Will update as it improves.
OK so I started having fun. Well, not so at the beginning. My setup is:
Intel Core 2 Duo E6300
P5W-DH Deluxe
Xtreme DDR2-667 1G x 2 (good up to DDR2-1080 / 4-4-3)
GeForce 6200TC
CNPS 9500
I booted the system without touching the BIOS other than disabling onboard sh*ts, etc., and raised the clocks little by little. (more on this later) So settings usually related to overclocking are at 'stock' configuration. (VCore, VDIMM, etc.) Every 50~100MHz I run either dual SP 32M or 5 mins of dual Prime95, then go up another notch. Findings:
1. Temps are extremely low @stock VCore. On this board, Pentium D 930 @4GHz used to hit 70C easily on load, and I've personally verified it with a thermo-gun. 😀 My E6300's idle temp basically equals to ambient temperature. (24C) This is a never-before-seen phenomenon to me.
2. I've stopped raising clocks @2.4GHz for now and currently running dual-Prime95 for about 40 mins. CPU Temp = 43C
http://img157.imageshack.us/my.php?image=2400apt0.png
http://img157.imageshack.us/my.php?image=2400btd8.png
3. The major obstacle so far has been the ASUS board. (What a surprise) At first I thought ASUS finally gave us a DFI-like board, but boy, was I wrong. You can still expect the usual ASUS behaviors from this board. Sure it's much improved (shouldn't have?) but expect to go through voltage fluctuation, warm-boot, cold-boot, having to raise FSB gradually, doing things on its own, etc. Again, it's much, much improved upon, say, A8N32-SLI, but it's still there and it's simply annoying.
4. At least ASUS implemented great BIOS recovery. I haven't had to clear CMOS yet.
Does anyone know what "FSB Termination Voltage" is and what it does, in the Award BIOS? Please post or PM me. I'd much appreciate it.
-------------------------------------
Edit 2: Super PI 32M @2.5GHz / Stock VCore
http://img195.imageshack.us/my.php?image=2500rf1.png
------------------------------------
Edit 3: Dual-Prime95 currently running @2.8GHz (7x400), STILL STOCK VCORE
http://img97.imageshack.us/my.php?image=2800aft0.png
-----------------------------------
Edit 4: Going to bed leaving the system running 2 instances of Prime95 @2.94GHz. Will report back tomorrow.
Edit 5: I'd consider the following stable.
http://img225.imageshack.us/my.php?image=2940prime4hrs3ed1.png
Also I could raise the FSB up to 430, but that required running memory with speed less than 1:1, which should be avoided on Intel platform. When I have more time, I will run the Prime95 again just to make sure.
Basically it's my first experience of motherboard (FSB, to be exact) limiting a CPU OC'ing. For the above Large FFT run, I had given 2.25VDIMM for memory. Had I given 2.30V, it'd have run much longer, is my guess. What happens is the northbridge is stressed by high FSB and memory (frequency/timing both), so even though CPU and memory have more potential, I won't see it on this board. Not that this board is bad in any stretch. This is the best ASUS board I've ever purchased, by far. I will make it short and easy:
1. E6300 is just like Socket 939 Opterons. 50% OC'ing seems easy.
2. Your motherboard's capability of high FSB will be the most limiting factor in OC'ing this low-multiplier CPU.
3. MCH (northbridge) gets stressed by both high FSB and memory.
4. This results in another problem - What if you won't be able to fully utilize your expensive memory because the memory controller on MCH can't handle it? This is a similar situation as in A64 OC'ing, where often times IMC used to be limiting memory performance. My sticks could do DDR2-800 @2.2V and DDR2-1080 @2.3 when FSB was under 300.
http://img95.imageshack.us/my.php?image=ddr21080ig9.png
But with the kind of FSB required for E6300 OC'ing, the story changes. I needed 2.3V for DDR2-840, and also had to loosen the timings. With anything higher, system just wouldn't boot.
5. On a side note, I'd strongly suggest an aftermarket cooler. It could be an isolated case but the IHS on my E6300 was very rough and uneven.
6. E6300 runs very very cool. Granted my system is an open-air system, the load temp was around 50C during dual-priming @2.94GHz.
7. P5W-DH users, make sure to disable "Hyperpath 3" when overclocking. I learned it in a hard way.
Now it's a benching time!
Small update: I'm currently testing @3010MHz (7x430) with dual-Prime95 Small FFT. While I'm happy the board is capable of doing 430HTT, it doesn't really mean much performance wise since I had to lower the memory speed below the FSB. (4:3) Will link to a pic after the Prime continues for at least a couple hours. VCore = 1.35V
-----------------------------------
OK here you go.
http://img220.imageshack.us/my.php?image=3010smallfft1nn5.png
http://img220.imageshack.us/my.php?image=3010smallfft2fr8.png
-----------------------------------
LOL once I gave up 1:1 the FSB goes up higher.. Booted @7x440. FSB:Memory=4:3
http://i35.photobucket.com/albums/d177/PenguinBelly/3080SmallFFTstart.png
SOMEONE PLEASE LINK ME TO A VMCH MOD TUTORIAL!
