![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
For those that don't like to read I have a 4.2ghz stable OC with 1.29v idle, 1.264v load Vcore, 1.62v Vdimm, 1.34QPI, 1.24IOH. Core i7 920 D0 (SLBEJ-3903A468). Is 1.34v for the QPI going to greatly accelerate the rate in which my chip degrades?
Realtemp reports my Cores at getting no warmer than 70C most of the time. Intel states ~68C Tcase is safe for 24/7 use without accelerating chip degration so that would mean my cores could run 24/7 at ~73C assuming Cores are running 5C over Tcase.
On the other hand any monitoring utility reports my Tcase at 54-58C while folding. If I go by this reading my cores are only getting up to 63C at the most so maybe I need to calibrate the temps using the Core i7/Core2 calibration guide over at THG I have seen before (it appears to have been moved or deleted, my bookmarked link does not work
)
I had dug up This article over on anandtech about degration for a friend interested in how far he can push his chip and it got me thinking (again) about how much life I will get out of my i7 920. So, is 1.34v to the QPI going to be safe for 3-5 years?
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
For some more words on my experience with this chip you can continue reading:
I believe I lucked out and got one of the lower voltage chips (binned for market demand and not binned for the clocks achieved with a predefined voltage). Stock clocks under load Vcore is 1.05v, 1.15 with HT on using the motherboards Auto/Normal setting. 1.18v @ 3.8, 1.24@4.0, 1.264@4.2, 1.3@4.4.
QPI on the other had requires 1.2v-1.24v to run my ram with tight timings or fast clocks with everything else at stock speeds.
My only concern with degration is using 1.34v for the QPI. Even 3.8ghz requires 1.32v which isn't much lower than the 1.34v i need for 4.2.
My last system was a Pentium 4 2.4c running at 3.5ghz with 1.58v (1.55 idle 1.55 load-droop modded motherboard) for 6 years. It saw some time folding and a lot of gaming and encoding. It still runs just fine, although I had a couple of crashes/lockups over the last month I was using it before I built this new rig.
Current settings. Bold voltages are bios followed by what I see in a monitoring program
4200mhz. 200x21
1.325v -- 1.29v idle, 1.264v under load (LLC is off)
1.8v CPU PLL
1.34v QPI , DDR1600 requires 1.42v
1.2v QPI-PLL
1.62 vDIMM -- 1.616v for DDR1200 C6, 1.67v -- 1.648v DDR1600 C8
1.24v IOH
1.14v ICH
These settings are stable for 48 hours Prime 95 Small FFT + 48 hrs Prime Large FFT + 24hours mix + hundreds of ITB runs without rebooting,and no programs crash/hang/error. The memory is fine for 12 hours of Memtest86 (forgot how many passes I did).
Temps.
Room temperature is between 20C -- 23C ( 68F and 74F).
Idle:
CPU@35-40C
Cores@35-40C
NorthBridge@26-30C
Motherboard@26-30C
Folding (Prime95 gets ~4C warmer)
CPU@54-58C
Cores@61-66cool, 67-72warm
NorthBridge@43-48C
Motherboard@38C-42C
Intel says that the core temps should be ~5C warmer than the CPU TCASE temp and since I am seeing a difference of ~10C on average I am assuming that my motherboard is reading something incorrectly, but I feel that if the cores running at <70C most of the time is fairly safe temperature wise. Would it be worth calibrating RealTemp to get the average core temp to 5C over TCase???
The only concern I have is that my QPI is running at 1.34v, and 1.35 is the 'safe' maximum, and my
Realtemp reports my Cores at getting no warmer than 70C most of the time. Intel states ~68C Tcase is safe for 24/7 use without accelerating chip degration so that would mean my cores could run 24/7 at ~73C assuming Cores are running 5C over Tcase.
On the other hand any monitoring utility reports my Tcase at 54-58C while folding. If I go by this reading my cores are only getting up to 63C at the most so maybe I need to calibrate the temps using the Core i7/Core2 calibration guide over at THG I have seen before (it appears to have been moved or deleted, my bookmarked link does not work
)I had dug up This article over on anandtech about degration for a friend interested in how far he can push his chip and it got me thinking (again) about how much life I will get out of my i7 920. So, is 1.34v to the QPI going to be safe for 3-5 years?
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
For some more words on my experience with this chip you can continue reading:
I believe I lucked out and got one of the lower voltage chips (binned for market demand and not binned for the clocks achieved with a predefined voltage). Stock clocks under load Vcore is 1.05v, 1.15 with HT on using the motherboards Auto/Normal setting. 1.18v @ 3.8, 1.24@4.0, 1.264@4.2, 1.3@4.4.
QPI on the other had requires 1.2v-1.24v to run my ram with tight timings or fast clocks with everything else at stock speeds.
My only concern with degration is using 1.34v for the QPI. Even 3.8ghz requires 1.32v which isn't much lower than the 1.34v i need for 4.2.
My last system was a Pentium 4 2.4c running at 3.5ghz with 1.58v (1.55 idle 1.55 load-droop modded motherboard) for 6 years. It saw some time folding and a lot of gaming and encoding. It still runs just fine, although I had a couple of crashes/lockups over the last month I was using it before I built this new rig.
Current settings. Bold voltages are bios followed by what I see in a monitoring program
4200mhz. 200x21
1.325v -- 1.29v idle, 1.264v under load (LLC is off)
1.8v CPU PLL
1.34v QPI , DDR1600 requires 1.42v
1.2v QPI-PLL
1.62 vDIMM -- 1.616v for DDR1200 C6, 1.67v -- 1.648v DDR1600 C8
1.24v IOH
1.14v ICH
These settings are stable for 48 hours Prime 95 Small FFT + 48 hrs Prime Large FFT + 24hours mix + hundreds of ITB runs without rebooting,and no programs crash/hang/error. The memory is fine for 12 hours of Memtest86 (forgot how many passes I did).
Temps.
Room temperature is between 20C -- 23C ( 68F and 74F).
Idle:
CPU@35-40C
Cores@35-40C
NorthBridge@26-30C
Motherboard@26-30C
Folding (Prime95 gets ~4C warmer)
CPU@54-58C
Cores@61-66cool, 67-72warm
NorthBridge@43-48C
Motherboard@38C-42C
Intel says that the core temps should be ~5C warmer than the CPU TCASE temp and since I am seeing a difference of ~10C on average I am assuming that my motherboard is reading something incorrectly, but I feel that if the cores running at <70C most of the time is fairly safe temperature wise. Would it be worth calibrating RealTemp to get the average core temp to 5C over TCase???
The only concern I have is that my QPI is running at 1.34v, and 1.35 is the 'safe' maximum, and my
Last edited: