F5F Stay Refreshed Hardware Desktop Poor results with the brand new 3900x model.

Poor results with the brand new 3900x model.

Poor results with the brand new 3900x model.

Pages (3): Previous 1 2 3 Next
Y
YouriiruoY
Member
212
01-11-2020, 09:34 AM
#11
I believe the settings exceed what's allowed in game mode. The R20 rating is quite low. A few adjustments are needed too. Even with game mode active, the score should hover near 3800 and creator mode close to 7400 using default (no PBO, auto OC).
Y
YouriiruoY
01-11-2020, 09:34 AM #11

I believe the settings exceed what's allowed in game mode. The R20 rating is quite low. A few adjustments are needed too. Even with game mode active, the score should hover near 3800 and creator mode close to 7400 using default (no PBO, auto OC).

B
Blue_n0se
Member
57
01-11-2020, 10:05 AM
#12
Based on my search and testing, results hover around 6900-7200. If you're at 7400 with stock settings, that's a great achievement—looks like you have a solid sample. I'm curious about how memory type and capacity affect performance.
B
Blue_n0se
01-11-2020, 10:05 AM #12

Based on my search and testing, results hover around 6900-7200. If you're at 7400 with stock settings, that's a great achievement—looks like you have a solid sample. I'm curious about how memory type and capacity affect performance.

D
DrPingouin
Member
214
01-11-2020, 04:36 PM
#13
I possess a small lemon. My chip requires 1.368v at 4.3Ghz for stable performance. My R20 rating at 4.3Ghz is 7710, which is typical for the default setting. At baseline, the CPU will increase to around 4.1Ghz across all cores. A key factor is temperature—higher temps cause lower boosts. I have a 280mm AIO, nothing unusual. The peak temperature reaches 72°C when using the default AIDA64 setup. Lower scores you see online likely reflect systems with stock coolers or less efficient cooling methods.
D
DrPingouin
01-11-2020, 04:36 PM #13

I possess a small lemon. My chip requires 1.368v at 4.3Ghz for stable performance. My R20 rating at 4.3Ghz is 7710, which is typical for the default setting. At baseline, the CPU will increase to around 4.1Ghz across all cores. A key factor is temperature—higher temps cause lower boosts. I have a 280mm AIO, nothing unusual. The peak temperature reaches 72°C when using the default AIDA64 setup. Lower scores you see online likely reflect systems with stock coolers or less efficient cooling methods.

X
xLillyLOL
Junior Member
25
01-18-2020, 06:47 PM
#14
Reviewing 3900x with stock configurations, XMP disabled. Current idle temps range 7040-7100, load temps around 68°C with 2x 360 rad peak speeds. Peak performance hits 4140 across all cores. Single-core results show 4385. Numbers reflect AMD Ryzen Master, latest BIOS (10F?). When testing real applications, spikes reach 46xx on some cores. Previous ABBA BIOS/7C versions differ. Hope this helps as a reference.
X
xLillyLOL
01-18-2020, 06:47 PM #14

Reviewing 3900x with stock configurations, XMP disabled. Current idle temps range 7040-7100, load temps around 68°C with 2x 360 rad peak speeds. Peak performance hits 4140 across all cores. Single-core results show 4385. Numbers reflect AMD Ryzen Master, latest BIOS (10F?). When testing real applications, spikes reach 46xx on some cores. Previous ABBA BIOS/7C versions differ. Hope this helps as a reference.

M
Monoki06
Member
152
01-18-2020, 07:30 PM
#15
It seems there is quite a variant amount chips. I remember many 3900x reviews have R15 scores between 3100- 3200. However Gamer Nexus hits 327x with auto OC in its review. Mine can do 3279 also with auto OC. Or 3244 without. My Cinebench scores remain more or less the same with old and new AGESA. I think RAM does nothing for Cinebench score. I have 3200MHz CL16 32GB. Either way, OP only scores 6446 is way too low for whatever reasons.
M
Monoki06
01-18-2020, 07:30 PM #15

It seems there is quite a variant amount chips. I remember many 3900x reviews have R15 scores between 3100- 3200. However Gamer Nexus hits 327x with auto OC in its review. Mine can do 3279 also with auto OC. Or 3244 without. My Cinebench scores remain more or less the same with old and new AGESA. I think RAM does nothing for Cinebench score. I have 3200MHz CL16 32GB. Either way, OP only scores 6446 is way too low for whatever reasons.

S
Selini
Member
161
01-20-2020, 01:01 AM
#16
It was a competitive setup. But you're correct, my readings are quite high. Shouldn't be overlooking the custom loop with 2 360 mm rads and six fans. At maximum I reached 78 degrees. I used Kryonaut paste and I'm confident in my application (I've done it many times). Voltages are normal at 1.4v most of the time. I need to adjust them down to lower temperatures and hopefully boost my scores. Do you have any guides or references? I'm not a big overclocker, mostly working on the Intel side which was surprisingly easy.
S
Selini
01-20-2020, 01:01 AM #16

It was a competitive setup. But you're correct, my readings are quite high. Shouldn't be overlooking the custom loop with 2 360 mm rads and six fans. At maximum I reached 78 degrees. I used Kryonaut paste and I'm confident in my application (I've done it many times). Voltages are normal at 1.4v most of the time. I need to adjust them down to lower temperatures and hopefully boost my scores. Do you have any guides or references? I'm not a big overclocker, mostly working on the Intel side which was surprisingly easy.

W
Whift
Junior Member
8
01-21-2020, 04:44 AM
#17
Find the exact fan owners and review the specifications. In BIOS, navigate to the fan settings and verify that the fans on both radiators cool based on the CPU sensor, not the system sensor or any other. Review the fan curve; this is the setting I apply to all fans (2x 3-way splitter), and the pump closely follows the same curve. Seek a balance that suits your needs. Note: At full load, fans reach 2000 RPM—why this value seems off is unclear.
W
Whift
01-21-2020, 04:44 AM #17

Find the exact fan owners and review the specifications. In BIOS, navigate to the fan settings and verify that the fans on both radiators cool based on the CPU sensor, not the system sensor or any other. Review the fan curve; this is the setting I apply to all fans (2x 3-way splitter), and the pump closely follows the same curve. Seek a balance that suits your needs. Note: At full load, fans reach 2000 RPM—why this value seems off is unclear.

V
vwgti2a
Member
106
01-24-2020, 07:38 PM
#18
I see Ryzen Master and HWinfo giving varying Vcore readings. The BIOS usually sets it to 1.225v, but HWinfo often shows around 1.226v. Ryzen Master tends to be slightly higher, between 0.06 and 0.1v. During Cinebench R20 it stays near 1.32 to 1.34v. When adjusting with undervolting or overvolting, a -0.025v setting usually yields the best performance, boosting scores by about 15 points. Going too far under or overvolting can cause a significant drop in scores. I use a standard 280mm AIO with two Noctua fans, keeping the fan speed at 1000rpm for quiet operation. At default settings it reaches about 72°C, or 84°C if I push to 4.3GHz manually. It seems your motherboard might be setting the Vcore too high, which affects both low scores and overheating.
V
vwgti2a
01-24-2020, 07:38 PM #18

I see Ryzen Master and HWinfo giving varying Vcore readings. The BIOS usually sets it to 1.225v, but HWinfo often shows around 1.226v. Ryzen Master tends to be slightly higher, between 0.06 and 0.1v. During Cinebench R20 it stays near 1.32 to 1.34v. When adjusting with undervolting or overvolting, a -0.025v setting usually yields the best performance, boosting scores by about 15 points. Going too far under or overvolting can cause a significant drop in scores. I use a standard 280mm AIO with two Noctua fans, keeping the fan speed at 1000rpm for quiet operation. At default settings it reaches about 72°C, or 84°C if I push to 4.3GHz manually. It seems your motherboard might be setting the Vcore too high, which affects both low scores and overheating.

V
Viizion_PvPz
Senior Member
670
01-24-2020, 07:44 PM
#19
I appreciate my fans: Corsair LL120 I find it actually appealing based on water temperature, not just CPU heat. Having the fans increase whenever I launch an app feels pointless and irritating because brief spikes don’t really raise water temps much over time. It’s the cooling system that needs attention, at least. That’s my take, and it matches what most in the water cooling community say. My setup is a curve: my RAM stuck at 1866, XMP profile used, Cinebench spiked to 6999. I also cut the CPU voltage to 1.3v, and now idle stays steady at 41°C, full load at 57°C. I’d say the problem is mostly solved, haha. Ryzen seems more sensitive than Intel did back then. When Skylake arrived (my old CPU), it had its issues, but nothing this strange. Reaching 81°C at base speed unless voltages drop is just not acceptable, regardless of AMD’s claims. No wonder AIO users hit their limits even with light overclocks. I’m skipping the overclocking now since it looks like you can’t gain much no matter what. I just set a PBO overclock in Ryzen Master and call it a day. Thanks for all the support!
V
Viizion_PvPz
01-24-2020, 07:44 PM #19

I appreciate my fans: Corsair LL120 I find it actually appealing based on water temperature, not just CPU heat. Having the fans increase whenever I launch an app feels pointless and irritating because brief spikes don’t really raise water temps much over time. It’s the cooling system that needs attention, at least. That’s my take, and it matches what most in the water cooling community say. My setup is a curve: my RAM stuck at 1866, XMP profile used, Cinebench spiked to 6999. I also cut the CPU voltage to 1.3v, and now idle stays steady at 41°C, full load at 57°C. I’d say the problem is mostly solved, haha. Ryzen seems more sensitive than Intel did back then. When Skylake arrived (my old CPU), it had its issues, but nothing this strange. Reaching 81°C at base speed unless voltages drop is just not acceptable, regardless of AMD’s claims. No wonder AIO users hit their limits even with light overclocks. I’m skipping the overclocking now since it looks like you can’t gain much no matter what. I just set a PBO overclock in Ryzen Master and call it a day. Thanks for all the support!

S
swagdude9787
Member
81
01-24-2020, 09:19 PM
#20
It seems the issue lies more with MBOO vendors adjusting the Vcore too high, not with AMD itself. Since AGESA ABBA, those voltage spikes under light loads have been resolved.
S
swagdude9787
01-24-2020, 09:19 PM #20

It seems the issue lies more with MBOO vendors adjusting the Vcore too high, not with AMD itself. Since AGESA ABBA, those voltage spikes under light loads have been resolved.

Pages (3): Previous 1 2 3 Next