Look into how the Static Read Control works(its also called other things).
From what I remember it works like memory timings but for the northbridge memory controller but it will have an affect on the limits of the fsb.

I've heard that disabling static read control is recommended for high overclocks, though I've tried it alongside looser ram timings, and it doesn't seem to help in my case. I may give it another try, but with the current timings I use now. Then again, it could just be because I was initially targeting 4GHz with a much higher FSB and a reduced multiplier. My goal is to reach a 400MHz FSB, as this is the fastest my chipset can go before hitting the next speed strap. Past 400MHz, it goes to a strap with looser internal timings, and the RAM actually slows down as a result, until you reach ~466MHz FSB. My board will boot with a 460MHz FSB, but it'll throw errors left and right in Memtest86+.

EDIT: It turns out my machine wasn't mprime-stable with the original settings I had. To fix this, I decided to bump the RAM voltage to 2.00v, the FSB termination voltage to 1.450v, and the northbridge voltage to 1.7v. So far, it seems to be working. 😉

Perhaps you have already tried this.

DRAM RAS# to CAS# Delay: 4 Change to 5
DRAM RAS# Precharge: 3 Change to 4

Be sure that "Command per clock" also known as CPC is turned off if you have this setting. The setting can also be 1T/2T instead of on/off then you want 2T.

The best is perhaps to use really loose ram timings all over when trying to get a higher overclock and try to get the memory back up to speed after you have found a higher stable CPU freq.
You should have a good chance of reaching 4GHz with an E8500 even on i965 with 8GB memory but higher will be difficult.

^I could try that. I can't change my command rate, but it's at 2T regardless.

Alright, I think I've managed to get it stable at 400MHz FSB with the memory timings you've recommended, as well as a ram voltage bump to 2.1v, and a couple of vcore bumps (~1.1675v right now, with the actual value reading as 1.14, due to vdroop). I'll have to let mprime run a bit longer, and maybe run it by itself without any other programs just to see.

EDIT: Now i'm going to see if I can get this thing to 4GHz, and if the strap change will make any actual difference in memory performance. 😎

I'm using a P5B on my main rig with a Xeon E5430 via the 771 -> 775 hack. The BIOS has been updated to include the latest OROMs and microcodes.

My experience has generally been good witht his board. Having said that, I knew what I was in for when I bought it as is. These boards need complete re-caps before they should be used in a productive environment. The VRM consists of defective UCC KZG series capacitors on the VRM high, and electrolytic 6mm Panasonic FL on the VRM low. Panasonic FL age quickly, like most electrolytic low-ESR caps. The VRM high caps should be replaced by Nichicon HM class caps or better, or 8mm 16V 270uF (or higher) polymer caps.

The VRM low should be replaced by 6x6 or 6x8 polymer capacitors. UCC PSC 2.5V 820uF is an excellent and inexpensive replacement. There is one extra spot in the VRM high and as well as in the VRM low which Asus did not place a cap. They should be added to enhance the overall ripple suppression.

The board is littered with Toshin Kyogo 6.3V 820uF capacitors which always fail, especially the one near the southbridge. They should all be replaced with Nichicon HM class or better. There are two polymer caps near the southbridge which should be left alone. These are the yellow/black and FPCap/PSC.

The on board audio outputs unecessarily faintly, but I'm sure this can be remedied by hacking the driver. The latest audio driver can be downloaded from Intel, but it must be hacked to include the Dev ID of the Asus variant.

The onboard JMB363 is useless for burning and should not be used other than for HDDs. The back eSata port which is connected to the JMB363 works perfectly. If you need to use an IDE Optical drive, get a SIL0680 PCI controller, or attach an IDE-SATA converter to the back of the drive.

The board advertises FSB1333 capability, but in my experience, the P965 isn't designed for anything over 1066 and will overheat. I run my chip at 1066 which gives me an overall clock speed of 2.1Ghz instead of 2.66.

If you want the best performance out of this platform, my advice to you is to purchase a Xeon E5450 and run it at 1066 which will give you 2.4Ghz. Stay away from 120W Xeons (and the 1600Mhz FSB ones, for that matter). The 80W chips run hot enough as it is (Mine runs at 45C-55C idle and reaches about 60 under very heavy load).

The board has absolutely no trouble at all with 8GB of RAM. I've got 4 x 2GB modules of Micron RAM, and it was recognized perfectly. Just make sure to enable memory re-mapping in the BIOS if you want your OS to see everything.

All-in-all, an excellent board from 2008. Perhaps the next best board since the venerable P2B. Unlike the P2B however, it was held down with shoddy capacitors.

If you want my modified/updated BIOS, LMK.

mr_bigmouth_502 wrote:

Alright, I think I've managed to get it stable at 400MHz FSB with the memory timings you've recommended, as well as a ram voltage bump to 2.1v, and a couple of vcore bumps (~1.1675v right now, with the actual value reading as 1.14, due to vdroop). I'll have to let mprime run a bit longer, and maybe run it by itself without any other programs just to see.

EDIT: Now i'm going to see if I can get this thing to 4GHz, and if the strap change will make any actual difference in memory performance. 😎

I think you need more Vcore to hit 4GHz and beyond. You seem rather conservative with Vcore. Max VID is 1.3625V as specified by Intel. And max Vcore you should feed a Wolfdale is 1.45V. Try setting Vcore at 1.3625V and see how far you can go.

Also, your FSB term voltage is rather high. You're at the Intel specified limit of 1.45V. That's probably only required near 500MHz FSB. You might want to lower it (1.30V?).

You should aim for the highest CPU multi, that'll reduce the FSB strain (and voltage) on your NB especially since you're clocking with 4 DIMMs populated. Instead of higher FSB and lower multi.

mockingbird wrote:

If you want the best performance out of this platform, my advice to you is to purchase a Xeon E5450 and run it at 1066 which will give you 2.4Ghz. Stay away from 120W Xeons (and the 1600Mhz FSB ones, for that matter). The 80W chips run hot enough as it is (Mine runs at 45C-55C idle and reaches about 60 under very heavy load).

Funny, Wolfdale/Yorkfield was supposed to run pretty cool. My E5430 runs a bit too cool in fact, due to a faulty temp sensor that reads slightly below ambient (around 28-30 deg C) when idle. I've undervolted mine to 0.9750V but even before that, I don't remember such high temps at idle. That said, I'm using a Antec 620 CLC on my chip.

mockingbird wrote:

I'm using a P5B on my main rig with a Xeon E5430 via the 771 -> 775 hack. The BIOS has been updated to include the latest OROMs […]
Show full quote

I'm using a P5B on my main rig with a Xeon E5430 via the 771 -> 775 hack. The BIOS has been updated to include the latest OROMs and microcodes.

My experience has generally been good witht his board. Having said that, I knew what I was in for when I bought it as is. These boards need complete re-caps before they should be used in a productive environment. The VRM consists of defective UCC KZG series capacitors on the VRM high, and electrolytic 6mm Panasonic FL on the VRM low. Panasonic FL age quickly, like most electrolytic low-ESR caps. The VRM high caps should be replaced by Nichicon HM class caps or better, or 8mm 16V 270uF (or higher) polymer caps.

The VRM low should be replaced by 6x6 or 6x8 polymer capacitors. UCC PSC 2.5V 820uF is an excellent and inexpensive replacement. There is one extra spot in the VRM high and as well as in the VRM low which Asus did not place a cap. They should be added to enhance the overall ripple suppression.

The board is littered with Toshin Kyogo 6.3V 820uF capacitors which always fail, especially the one near the southbridge. They should all be replaced with Nichicon HM class or better. There are two polymer caps near the southbridge which should be left alone. These are the yellow/black and FPCap/PSC.

The on board audio outputs unecessarily faintly, but I'm sure this can be remedied by hacking the driver. The latest audio driver can be downloaded from Intel, but it must be hacked to include the Dev ID of the Asus variant.

The onboard JMB363 is useless for burning and should not be used other than for HDDs. The back eSata port which is connected to the JMB363 works perfectly. If you need to use an IDE Optical drive, get a SIL0680 PCI controller, or attach an IDE-SATA converter to the back of the drive.

The board advertises FSB1333 capability, but in my experience, the P965 isn't designed for anything over 1066 and will overheat. I run my chip at 1066 which gives me an overall clock speed of 2.1Ghz instead of 2.66.

If you want the best performance out of this platform, my advice to you is to purchase a Xeon E5450 and run it at 1066 which will give you 2.4Ghz. Stay away from 120W Xeons (and the 1600Mhz FSB ones, for that matter). The 80W chips run hot enough as it is (Mine runs at 45C-55C idle and reaches about 60 under very heavy load).

The board has absolutely no trouble at all with 8GB of RAM. I've got 4 x 2GB modules of Micron RAM, and it was recognized perfectly. Just make sure to enable memory re-mapping in the BIOS if you want your OS to see everything.

All-in-all, an excellent board from 2008. Perhaps the next best board since the venerable P2B. Unlike the P2B however, it was held down with shoddy capacitors.

If you want my modified/updated BIOS, LMK.

My experience with P965 is different. I have reached 500 (2000) MHz FSB with some old Gigabyte P965 DS3 board, that board did have fresh caps though.

Core 2 Duo 45nm can handle up to ~1.5V core voltage (with good cooling) but when it comes to CPU VTT be careful. I agree with Zenn, do not go higher than 1.45V CPU VTT.

Zenn wrote:

Funny, Wolfdale/Yorkfield was supposed to run pretty cool. My E5430 runs a bit too cool in fact, due to a faulty temp sensor that reads slightly below ambient (around 28-30 deg C) when idle. I've undervolted mine to 0.9750V but even before that, I don't remember such high temps at idle. That said, I'm using a Antec 620 CLC on my chip.

Is your 5430 C0 or E0? That could have something to do with it.

What are your individual core temps?

I'm tempted to try undervolting. I've got to put up with the vdroop on this board though. My MB sensor was haywire when I first powered the thing, reporting 30-40C over the real temperature. It seems to have settle down now.

Skyscraper wrote:

My experience with P965 is different. I have reached 500 (2000) MHz FSB with some old Gigabyte P965 DS3 board, that board did have fresh caps though.

Core 2 Duo 45nm can handle up to ~1.5V core voltage (with good cooling) but when it comes to CPU VTT be careful. I agree with Zenn, do not go higher than 1.45V CPU VTT.

I'm not saying it's not capable of sustaining these speeds, but for how long? I don't like my northbridges to be so hot that I can't touch them and hold my finger there for 5 seconds. You run your northbridge past what it was designed for (And Asus can advertise it as 1333FSB capable if they want, it's still not a 1333FSB board), and it will fail prematurely. Heck, even the P35/P45 run too hot for my liking, and I added a small fan to my G45 Core2Quad board because of this. I guess I might experiment with it if the Asus northbridge heatsink were better suited to accommodate a fan, but the top is shaped in such a way so that active cooling wouldn't help it.

I should really dabble with undervolting though.

ZENN, is your undervolted E5430 Prime95 stable on all four cores? I'm using the stock copper-core Intel cooler off my Pentium D 925. It definitely runs a lot cooler than the Pentium D.

mockingbird wrote:

I'm using a P5B on my main rig with a Xeon E5430 via the 771 -> 775 hack. The BIOS has been updated to include the latest OROMs […]
Show full quote

I'm using a P5B on my main rig with a Xeon E5430 via the 771 -> 775 hack. The BIOS has been updated to include the latest OROMs and microcodes.

My experience has generally been good witht his board. Having said that, I knew what I was in for when I bought it as is. These boards need complete re-caps before they should be used in a productive environment. The VRM consists of defective UCC KZG series capacitors on the VRM high, and electrolytic 6mm Panasonic FL on the VRM low. Panasonic FL age quickly, like most electrolytic low-ESR caps. The VRM high caps should be replaced by Nichicon HM class caps or better, or 8mm 16V 270uF (or higher) polymer caps.

The VRM low should be replaced by 6x6 or 6x8 polymer capacitors. UCC PSC 2.5V 820uF is an excellent and inexpensive replacement. There is one extra spot in the VRM high and as well as in the VRM low which Asus did not place a cap. They should be added to enhance the overall ripple suppression.

The board is littered with Toshin Kyogo 6.3V 820uF capacitors which always fail, especially the one near the southbridge. They should all be replaced with Nichicon HM class or better. There are two polymer caps near the southbridge which should be left alone. These are the yellow/black and FPCap/PSC.

The on board audio outputs unecessarily faintly, but I'm sure this can be remedied by hacking the driver. The latest audio driver can be downloaded from Intel, but it must be hacked to include the Dev ID of the Asus variant.

The onboard JMB363 is useless for burning and should not be used other than for HDDs. The back eSata port which is connected to the JMB363 works perfectly. If you need to use an IDE Optical drive, get a SIL0680 PCI controller, or attach an IDE-SATA converter to the back of the drive.

The board advertises FSB1333 capability, but in my experience, the P965 isn't designed for anything over 1066 and will overheat. I run my chip at 1066 which gives me an overall clock speed of 2.1Ghz instead of 2.66.

If you want the best performance out of this platform, my advice to you is to purchase a Xeon E5450 and run it at 1066 which will give you 2.4Ghz. Stay away from 120W Xeons (and the 1600Mhz FSB ones, for that matter). The 80W chips run hot enough as it is (Mine runs at 45C-55C idle and reaches about 60 under very heavy load).

The board has absolutely no trouble at all with 8GB of RAM. I've got 4 x 2GB modules of Micron RAM, and it was recognized perfectly. Just make sure to enable memory re-mapping in the BIOS if you want your OS to see everything.

All-in-all, an excellent board from 2008. Perhaps the next best board since the venerable P2B. Unlike the P2B however, it was held down with shoddy capacitors.

If you want my modified/updated BIOS, LMK.

I'll be honest, I'm TERRIBLE at soldering. 😜 If I'm going to go through the effort of recapping my board, I may as well save up a few dollars and go for a newer board. Right now, I have my sights set on an ASUS P5KC, since it has all the nice overclocking features that the P5B has, but with a newer chipset and DDR3 support. As well, it's a hybrid DDR2/DDR3 board, so I can upgrade my RAM later on. Then, when I finally decide to move on to a new platform, I'll already have the RAM I need. 😁

Ironically, I have my optical drive hooked to the JMB controller, and it works fine. 😜 It's actually the only way I was able to get it to work under Linux.

I have considered the Xeon mod, since you can get LGA 771 Xeons stupidly cheap on Ebay with the converters, and since they're normally sold in pairs for dual CPU boards, I'd be able to install a better CPU in both of my LGA775 boards. 😁

And thanks for the offer on the modded BIOS. 😁 I wonder if it'll help with my current OCing situation.

Zenn wrote:

I think you need more Vcore to hit 4GHz and beyond. You seem rather conservative with Vcore. Max VID is 1.3625V as specified by […]
Show full quote

mr_bigmouth_502 wrote:

Alright, I think I've managed to get it stable at 400MHz FSB with the memory timings you've recommended, as well as a ram voltage bump to 2.1v, and a couple of vcore bumps (~1.1675v right now, with the actual value reading as 1.14, due to vdroop). I'll have to let mprime run a bit longer, and maybe run it by itself without any other programs just to see.

EDIT: Now i'm going to see if I can get this thing to 4GHz, and if the strap change will make any actual difference in memory performance. 😎

I think you need more Vcore to hit 4GHz and beyond. You seem rather conservative with Vcore. Max VID is 1.3625V as specified by Intel. And max Vcore you should feed a Wolfdale is 1.45V. Try setting Vcore at 1.3625V and see how far you can go.

Also, your FSB term voltage is rather high. You're at the Intel specified limit of 1.45V. That's probably only required near 500MHz FSB. You might want to lower it (1.30V?).

You should aim for the highest CPU multi, that'll reduce the FSB strain (and voltage) on your NB especially since you're clocking with 4 DIMMs populated. Instead of higher FSB and lower multi.

The reason my Vcore is set so low is because I have a crappy cooler, and I'm trying to prevent my core temperatures from going over 65C. Oddly, I've noticed that the mobo CPU temperature sensor, the ATK0100, often reports temperatures 10 degrees above what the cores read. I don't know whether I should trust this measurement or ignore it, as it's what the BIOS uses for its CPU temperature readings.

I could probably lower my FSB termination voltage. The only reason I increased it is because I thought it would make things more stable. Also, I once noticed that raising it would increase my memory speed in Memtest86+. The memory speed ratings given by that program are a little suspect though, as I've seen them drop from bumping up the FSB, which makes absolutely no sense.

mr_bigmouth_502 wrote:

I'll be honest, I'm TERRIBLE at soldering. 😜 If I'm going to go through the effort of recapping my board, I may as well save up a few dollars and go for a newer board. Right now, I have my sights set on an ASUS P5KC, since it has all the nice overclocking features that the P5B has, but with a newer chipset and DDR3 support. As well, it's a hybrid DDR2/DDR3 board, so I can upgrade my RAM later on. Then, when I finally decide to move on to a new platform, I'll already have the RAM I need. 😁

Sounds good. I'd rather have 4 DDR3 slots than only two though.

Ironically, I have my optical drive hooked to the JMB controller, and it works fine. 😜 It's actually the only way I was able to get it to work under Linux.

Try burning a dual-layer DVD. incidentally, JMB is the only OROM I didn't update in the BIOS because it hangs in detection when you use the latest one with AHCI mode for the JMB controller.

I have considered the Xeon mod, since you can get LGA 771 Xeons stupidly cheap on Ebay with the converters, and since they're normally sold in pairs for dual CPU boards, I'd be able to install a better CPU in both of my LGA775 boards. 😁

You're probably better off with a really fast E84xx than with a quad-core.

And thanks for the offer on the modded BIOS. 😁 I wonder if it'll help with my current OCing situation.

Probably not, but here you go:

The BIOS also has another pleasant surprise added to it for those using Windows 7 and needing to SLICE 2 and 1.

I wish this board would allow less VCore. I put it down all the way and it tests stable with Prime95.

The P5KC will really be more of an in-between upgrade until I can afford to splurge on a nice i5 build. As it is, it would cost me far too much to upgrade to a whole new platform (CPU, mobo, RAM) all at once. The P5KC would allow me to squeeze more out of my current CPU and RAM.

I don't think I even have any dual layer DVDs around. 😜 My burner has always been a bit flaky though, so I use it sporadically, and I stick to burning at low speeds. It's pretty much only used for making OS discs, and ripping the occasional CD. Do they even make reliable optical drives anymore, or are they all just good for making coasters?

What are the downsides of using a quad-core on my current board, aside from the higher wattage?

Thank you for the link. 😁 I'll install it once I have this thing stable again. I don't want to be installing a corrupt BIOS. 😜

mr_bigmouth_502 wrote:

The P5KC will really be more of an in-between upgrade until I can afford to splurge on a nice i5 build. As it is, it would cost […]
Show full quote

The P5KC will really be more of an in-between upgrade until I can afford to splurge on a nice i5 build. As it is, it would cost me far too much to upgrade to a whole new platform (CPU, mobo, RAM) all at once. The P5KC would allow me to squeeze more out of my current CPU and RAM.

I don't think I even have any dual layer DVDs around. 😜 My burner has always been a bit flaky though, so I use it sporadically, and I stick to burning at low speeds. It's pretty much only used for making OS discs, and ripping the occasional CD. Do they even make reliable optical drives anymore, or are they all just good for making coasters?

What are the downsides of using a quad-core on my current board, aside from the higher wattage?

Thank you for the link. 😁 I'll install it once I have this thing stable again. I don't want to be installing a corrupt BIOS. 😜

Yea, the old parallel Lite-ON drives are great - once they're re-capped that is. I've got my DH20A4H with TDB firmware.

Lower clocks. Other than that, some stuff might benefit from the extra 2 cores and 12MB cache. IIRC, the lowly i3 trounced the multiple core AMD Bulldozer architecture in most games because games just aren't designed to take advantage of multiple (but albeit slower) cores. I reckon a 4Ghz E8x00 will do the same to a 775/771 quad core (Because of the old design and limited bandwidth).

What about emulators, like PCSX2 or Dolphin? They're coded to take advantage of multiple cores, though they're still heavily reliant on single-core performance.

Also, I'm running mprime right now, with the Vcore set to ~1.36v, and holy shit my temps are high. My core temperatures are at 84C right now, occasionally dropping to around 75C. The highest I saw was 85C. Am I risking permanently damaging my CPU here? Also, I had to crank the chipset termination voltage all the way back to 1.450V and the RAM voltage to 2.3v just to get it stable at my current settings. I think I'm going to turn down my Vcore a bit, and see what my other settings are set at just to double check.

My current settings are:

FSB: 416MHz
RAM Frequency: 832MHz
Memory Voltage: 2.3v
VCore: 1.35V (reduced from ~1.375); more like 1.30v under load due to vdroop
FSB Term.: 1.45v
NB Vcore: 1.70v
SB Vcore: 1.60v

CAS Latency: 4
RAS to CAS Delay: 5
RAS Precharge: 4
RAS Activate to Precharge: 13
Static Read Control: Faster

^Oddly, I find that if I disable Static Read Control and loosen my timings further than this, it actually creates more instabilities.

Right now, these settings seem fairly stable, and my core temps under load have reduced to 73C. Not ideal, but a hell of a lot better than 85C!

There are no dangers involved with running the Core 2 Duo at 85C, if the CPU degrades its from the FSB Term. voltage (CPU VTT) @ 1.45v not the heat. I would not worry too much but I would make sure the case is well ventilated so the rest of the componets dont get too hot and I would not run any Prime95 24h stability tests with the CPU VTT @ 1.45v. I would also add active cooling on the memory when running them @ 2.3V.

I felt my ram sticks and they were a bit warm, so I'm probably going to take some smallish fans I have in another case, and find a way to put them on without impeding the airflow of my CPU cooler. This is the cooler I'm using, a Rosewill RCX-Z775-EX. http://www.thinkcomputers.org/old/index.php?x … s&id=745&page=2 That may not be the exact model name, since I originally bought it for another system about 5 years ago, and I had to do some Googling until I found the model resembling mine.

I have a small fan from an old AMD CPU heatsink positioned in front of my northbridge heatsink, since it becomes too hot to touch without it. I also took the main front fan, and placed it in the side window, as all it was doing was keeping my hard drives cool. 🤣 I have another front fan, a 120mm, installed above my DVD drive. I ghetto-rigged a bit of a shroud on it out of some thin cardboard.

Anyway, why would the FSB termination Vcore do more to degrade the CPU than the CPU Vcore itself? I think I may try turning it down a notch to see if it remains stable. 😉

UPDATE: Alright, I just installed a fan to blow some air over my RAM sticks. I believe it's a 40 or 50mm fan, and it was originally one of the exhaust fans in an Antec Lanbox Lite case. It's not that great of a case, I remember the system that was in it ran quite hot, and the fans were LOUD when they were all working together. I have it positioned at the bottom of the RAM sticks, with my video card acting as a shelf. 🤣 The way it works out, it's right next to my northbridge fan! Maybe I'll swap that one for one of the Lanbox fans as well. 😉

mockingbird wrote:

Is your 5430 C0 or E0? That could have something to do with it.

What are your individual core temps?

It's an E0. But uhh I usually use Core Temp instead of HWMonitor. You can see they give conflicting temps. How I arrived at Core Temp's values was that from what I've read online, the E54xx TJMax is 85°C. So I used an offset of 15°C since the program's default is 100°C. FWIW I downloaded AIDA64 and it agrees with Core Temp. So... majority wins? 😉 Anyway my board is a Gigabyte EP45-UD3.

mockingbird wrote:

ZENN, is your undervolted E5430 Prime95 stable on all four cores? I'm using the stock copper-core Intel cooler off my Pentium D 925. It definitely runs a lot cooler than the Pentium D.

It's 2hrs Prime95 small fft stable. I know some people run longer but this is enough for me to test stability.

mr_bigmouth_502 wrote:

Anyway, why would the FSB termination Vcore do more to degrade the CPU than the CPU Vcore itself? I think I may try turning it down a notch to see if it remains stable. 😉

From Anandtech:

CPU VTT Voltage - VTT is the termination voltage for data lines used to interface the MCH with the CPU die(s) via the Front Side Bus. Higher values can provide additional FSB overclocking margin, especially with 45nm dual-core processors and quad-cores CPUs in general. We have never found any additional gains to be had above 1.36 when using air- or water-cooling. http://www.anandtech.com/show/2500/17

Also please see: Beware of VTT - it's a lurking time bomb... http://www.anandtech.com/show/3381

i ran my Pentium Dual Core for a long time with max VTT, and it still worked the last time I checked, but I had to dial down its OC to make it stable. Ahh, the joys of irresponsible overclocking. 🤣

Anyway, thanks for the heads up. 😉 I had been lusting after an E8500 for a long time, and now that I have one, I don't want to kill it.

UPDATE: I played around with the VTT some more, and I noticed my system was acting erratically, so I shut it down. I felt the RAM and it was hotter than before, so I suspected that that was the reason why. I originally thought that maybe the one fan wasn't providing enough cooling for the RAM, so I stacked them vertically to cover most of the length of the RAM. In the process, I discovered that the fan I originally installed for the RAM was never spinning in the first place, as part of the molex-to-fan converter was loose, so I replaced it. I booted my system up, new RAM cooling system in place, and my northbridge fan, which wasn't being held in place with anything, fell over for the umpteenth time and made an awful grinding noise, as usual. I said "to hell with it", took it out, and turned my RAM cooler on its side so one fan would blow on the northbridge heatsink, and the other would blow on the RAM. Now, my computer sounds like a jet engine, and I don't have to worry about my northbridge fan falling over, as it has a wider base. 🤣

Zenn wrote:

It's 2hrs Prime95 small fft stable. I know some people run longer but this is enough for me to test stability.

Interesting. I see your E0 has a lower power state that uses 7.5x multiplier. I believe the C0 only has two multiplier states - 8x and 6x.

Looking at SLBBK's feature notes, this might have something to do with it:

Notes on sSpec SLBBK ◾ OEM parts are available starting from Oct 6, 2008. ◾ Boxed processors are available starting from Aug 2 […]
Show full quote

Notes on sSpec SLBBK
◾ OEM parts are available starting from Oct 6, 2008.
◾ Boxed processors are available starting from Aug 29, 2008.
◾ Changes in E0 stepping: new XSAVE and XRSTORE instructions, ACNT2 feature used for determining processor utilization, halide-free package.

mr_bigmouth_502 wrote:

What about emulators, like PCSX2 or Dolphin? They're coded to take advantage of multiple cores, though they're still heavily reliant on single-core performance. PCSX2

Probably negligible. Proper multi-core support has only relatively recently begun being taken advantage of by developers (And was only enabled in GPU drivers a couple of years ago). Correct me if I'm wrong, but PCSX2 relies on GPU not CPU, so given the age of its core engine, I doubt you'd see much improvement with a quad core.