F5F Stay Refreshed Power Users Overclocking Issue with PHY bus caused by PhenomII x4 975BE mild OC

Issue with PHY bus caused by PhenomII x4 975BE mild OC

Issue with PHY bus caused by PhenomII x4 975BE mild OC

Pages (2): Previous 1 2
W
winndich
Member
156
10-02-2016, 09:12 AM
#11
Updated: CPU fan changed from 120x25 basic low noise PWM to 120x25 Delta 152 cfm full load burn temps fell to 49°C. Ran Ungine valley on max settings (GPU didn’t handle it well, limited frames a lot) – no artifacts detected, confirming my previous successful GPU overclock. No hangs or crashes observed. Tested stability and microflicks while playing games – everything fine so far. Also tried booting multiple Linux distributions from live USBs (Kali, Debian, Ubuntu) – no issues. Performance remains solid at 3.8. Planning to reboot and attempt 3.9 once posting is complete.
W
winndich
10-02-2016, 09:12 AM #11

Updated: CPU fan changed from 120x25 basic low noise PWM to 120x25 Delta 152 cfm full load burn temps fell to 49°C. Ran Ungine valley on max settings (GPU didn’t handle it well, limited frames a lot) – no artifacts detected, confirming my previous successful GPU overclock. No hangs or crashes observed. Tested stability and microflicks while playing games – everything fine so far. Also tried booting multiple Linux distributions from live USBs (Kali, Debian, Ubuntu) – no issues. Performance remains solid at 3.8. Planning to reboot and attempt 3.9 once posting is complete.

P
Prime3656
Member
158
10-09-2016, 09:25 AM
#12
Started Windows 3.9 with a minor vcore adjustment (1.365 up from 1.3500).
CPU-NB at 2.9, load temperature at 52°C.
Successfully completed overnight Prime95 test.
Small gain in Ungine Valley performance, still constrained by GPU.
Installed a 60mm fan to cool the VRM heatsinks directly; they are now noticeably cooler.
Planning to aim for 4.0 (3.0 CPU-NB).
UPDATE: 4.0 @1.365 vcore, 3.0 cpu-NB.
Load temperature dropped to 54°C.
Still needs overnight Prime95 verification to confirm stability, but it ran through OCCT and completed Ungine Valley.
VRM heatsinks are noticeably hotter than at 3.9 even with direct airflow; this is the maximum expected for the multi.
Will need to adjust using the hardware reference clock for fine tuning.
If stability is confirmed, I’ll keep the multi in place and slightly increase the reference clock.
P
Prime3656
10-09-2016, 09:25 AM #12

Started Windows 3.9 with a minor vcore adjustment (1.365 up from 1.3500).
CPU-NB at 2.9, load temperature at 52°C.
Successfully completed overnight Prime95 test.
Small gain in Ungine Valley performance, still constrained by GPU.
Installed a 60mm fan to cool the VRM heatsinks directly; they are now noticeably cooler.
Planning to aim for 4.0 (3.0 CPU-NB).
UPDATE: 4.0 @1.365 vcore, 3.0 cpu-NB.
Load temperature dropped to 54°C.
Still needs overnight Prime95 verification to confirm stability, but it ran through OCCT and completed Ungine Valley.
VRM heatsinks are noticeably hotter than at 3.9 even with direct airflow; this is the maximum expected for the multi.
Will need to adjust using the hardware reference clock for fine tuning.
If stability is confirmed, I’ll keep the multi in place and slightly increase the reference clock.

B
Bonnibel
Posting Freak
794
10-11-2016, 04:22 PM
#13
If that's stable, it makes a great clock for Phenom II. The voltage is quite low as well. I tested around a dozen Phenom II units and found they were all prone to leakage, needing more than 1.5v to reach 4GHz and running hot. At 3GHz it's a solid overclock on the NB, which is good because the L3 cache operates at that speed and cuts down RAM delays.
B
Bonnibel
10-11-2016, 04:22 PM #13

If that's stable, it makes a great clock for Phenom II. The voltage is quite low as well. I tested around a dozen Phenom II units and found they were all prone to leakage, needing more than 1.5v to reach 4GHz and running hot. At 3GHz it's a solid overclock on the NB, which is good because the L3 cache operates at that speed and cuts down RAM delays.

C
ChocksPvP
Junior Member
2
10-11-2016, 10:52 PM
#14
3.9 remains stable, still hoping for confirmation on 4.0. It's been running Prime95 now, and I'll check again tomorrow whether it holds up. So far it looks promising. The updated Phenom II model performs better than previous versions (up to the 965) in terms of voltage handling, while newer chips clock in at lower voltages. The 970, 975, and 980 chips are the latest revisions.
C
ChocksPvP
10-11-2016, 10:52 PM #14

3.9 remains stable, still hoping for confirmation on 4.0. It's been running Prime95 now, and I'll check again tomorrow whether it holds up. So far it looks promising. The updated Phenom II model performs better than previous versions (up to the 965) in terms of voltage handling, while newer chips clock in at lower voltages. The 970, 975, and 980 chips are the latest revisions.

T
TheMrMax0r
Junior Member
16
10-11-2016, 11:18 PM
#15
Completed overnight prime95 build and performed five full memtest runs to verify memory controller stability.
No issues detected during testing.
Adjusted voltage to 1.375V and raised HT ref clock to 203 for a 4060mhz CPU (4050 on my motherboard, which is slightly under the ref clock).
Operated at full load with temperature at 1.375V and 56°C.
If this setup remains stable, I’ll keep it. I’m nearing the cooling limits, and the VRM temperatures are also elevated.
No problems in Ungine Valley or Ungine Heaven; frame rate increased by about 1-2fps to compensate with a higher HT ref clock.
Completed OCCT for an hour without errors and am now running prime95.
T
TheMrMax0r
10-11-2016, 11:18 PM #15

Completed overnight prime95 build and performed five full memtest runs to verify memory controller stability.
No issues detected during testing.
Adjusted voltage to 1.375V and raised HT ref clock to 203 for a 4060mhz CPU (4050 on my motherboard, which is slightly under the ref clock).
Operated at full load with temperature at 1.375V and 56°C.
If this setup remains stable, I’ll keep it. I’m nearing the cooling limits, and the VRM temperatures are also elevated.
No problems in Ungine Valley or Ungine Heaven; frame rate increased by about 1-2fps to compensate with a higher HT ref clock.
Completed OCCT for an hour without errors and am now running prime95.

M
Melchio
Junior Member
22
10-12-2016, 05:28 PM
#16
That's a great overclock. You should tune your RAM frequency and timings next and find the most effective ratios.
M
Melchio
10-12-2016, 05:28 PM #16

That's a great overclock. You should tune your RAM frequency and timings next and find the most effective ratios.

C
Charliemc909
Posting Freak
898
10-13-2016, 02:21 AM
#17
Thank you. The rating looks solid at 1866 with a 9-9-9-24, which should provide some flexibility.
C
Charliemc909
10-13-2016, 02:21 AM #17

Thank you. The rating looks solid at 1866 with a 9-9-9-24, which should provide some flexibility.

G
GameBoosh
Senior Member
470
10-20-2016, 09:55 PM
#18
The maximum RAM capacity for stability is 2133, though it supports 1866. It boots Windows at 2400 with a minor voltage adjustment, but fails memtest unless timings are adjusted to 10-10-10-28. At 2133 with the same adjustment, it runs standard timings (9-9-9-24) and consistently passes memtest (I left it running overnight without any issues).

At 2133 with the same voltage tweak, it uses standard timings (9-9-9-24) and always succeeds in memtest (I allowed it to run continuously and never encountered a failure).

I’m using a Corsair memory cooling fan to maintain low temperatures. This caused a slight increase in frame rate during the ungine valley benchmark (GPU was the bottleneck). There’s minimal room for improvement on the GPU—besides possibly swapping it for a R9 390x, which I ordered today (concern about RX series compatibility is why I chose the R9). The Radeon 3650 is an older DDR2 model; even with added RAM sinks, it’s already running hot at 400MHz, so I didn’t increase its voltage. When I boosted the core to 750 using Radeon Bios Editor (tested in the Catalyst Control Center overdrive mode), performance was acceptable.

I used a budget aftermarket cooler (originally $15 on Amazon) and an Artic Silver Ceramic cooler. It didn’t significantly cool the card, which was expected since it was meant to be a basic 3D graphics card for Aero until I could get the R9 390x. Overall, it performs adequately for its price.

Choosing the 975BE CPU was more economical than the 980 and generally handled overclocking better than the Thuban core series (Phenom II X6). I also looked into the FX8350 because of its similar price to the 975, but after researching, I found that Windows 7 requires a hotfix for Bulldozer to function well, and the Phenom II models have higher instructions per clock than the FX line. Many games I play don’t fully utilize all four cores of the Phenom II, so adding more cores wouldn’t improve performance significantly.

Although newer technology would make the FX the superior choice, for my setup the Phenom II was the better fit.
G
GameBoosh
10-20-2016, 09:55 PM #18

The maximum RAM capacity for stability is 2133, though it supports 1866. It boots Windows at 2400 with a minor voltage adjustment, but fails memtest unless timings are adjusted to 10-10-10-28. At 2133 with the same adjustment, it runs standard timings (9-9-9-24) and consistently passes memtest (I left it running overnight without any issues).

At 2133 with the same voltage tweak, it uses standard timings (9-9-9-24) and always succeeds in memtest (I allowed it to run continuously and never encountered a failure).

I’m using a Corsair memory cooling fan to maintain low temperatures. This caused a slight increase in frame rate during the ungine valley benchmark (GPU was the bottleneck). There’s minimal room for improvement on the GPU—besides possibly swapping it for a R9 390x, which I ordered today (concern about RX series compatibility is why I chose the R9). The Radeon 3650 is an older DDR2 model; even with added RAM sinks, it’s already running hot at 400MHz, so I didn’t increase its voltage. When I boosted the core to 750 using Radeon Bios Editor (tested in the Catalyst Control Center overdrive mode), performance was acceptable.

I used a budget aftermarket cooler (originally $15 on Amazon) and an Artic Silver Ceramic cooler. It didn’t significantly cool the card, which was expected since it was meant to be a basic 3D graphics card for Aero until I could get the R9 390x. Overall, it performs adequately for its price.

Choosing the 975BE CPU was more economical than the 980 and generally handled overclocking better than the Thuban core series (Phenom II X6). I also looked into the FX8350 because of its similar price to the 975, but after researching, I found that Windows 7 requires a hotfix for Bulldozer to function well, and the Phenom II models have higher instructions per clock than the FX line. Many games I play don’t fully utilize all four cores of the Phenom II, so adding more cores wouldn’t improve performance significantly.

Although newer technology would make the FX the superior choice, for my setup the Phenom II was the better fit.

Pages (2): Previous 1 2