F5F Stay Refreshed Hardware Desktop 10850k weird bench is a curious project or item with a unique value.

10850k weird bench is a curious project or item with a unique value.

10850k weird bench is a curious project or item with a unique value.

Pages (3): Previous 1 2 3 Next
M
Marcustheduke
Senior Member
679
10-14-2020, 12:31 AM
#11
Benchmarking didn't go well even with some monitoring issues. On stock configurations, my scores were low—560 for single thread and 7002 for multi-thread in cpuz with 99 idle cycles. Temperatures were solid, around 60°C on multicore and 40°C on single. The monitor showed speeds between 4.9-5.1GHz on single core and 4.8-4.9GHz on multicore. The test manual locked all cores to 5.1/4.8 cache, resulting in 605 single and 7586 multi scores. It seems Gigabyte BIOS might struggle with boosting performance. Looking up this 4.9GHz model brought up many results. Overclocking manually fixes the problem, though it's a bit annoying.
M
Marcustheduke
10-14-2020, 12:31 AM #11

Benchmarking didn't go well even with some monitoring issues. On stock configurations, my scores were low—560 for single thread and 7002 for multi-thread in cpuz with 99 idle cycles. Temperatures were solid, around 60°C on multicore and 40°C on single. The monitor showed speeds between 4.9-5.1GHz on single core and 4.8-4.9GHz on multicore. The test manual locked all cores to 5.1/4.8 cache, resulting in 605 single and 7586 multi scores. It seems Gigabyte BIOS might struggle with boosting performance. Looking up this 4.9GHz model brought up many results. Overclocking manually fixes the problem, though it's a bit annoying.

S
Smoofie
Member
213
10-14-2020, 12:49 AM
#12
It seems your Gigabyte board didn't configure the CPU to match the default Intel specifications. A 560 single-thread score falls significantly short of expectations. Notably, a detailed CPU-Z Single Thread test reveals two threads are loading instead of one. The image displays the first task as fully loading one thread, while the second alternates between two cores, indicating the CPU primarily uses multipliers 51 and 52 during this test. Cinebench R23 is a genuine single-thread benchmark, loading just one thread. The screenshot shows an average multiplier of 52.74, meaning the CPU frequently switched to multiplier 53 (74% of time) and 52 (26% of time). This clearly confirms the presence and usage of the 53 multiplier.
S
Smoofie
10-14-2020, 12:49 AM #12

It seems your Gigabyte board didn't configure the CPU to match the default Intel specifications. A 560 single-thread score falls significantly short of expectations. Notably, a detailed CPU-Z Single Thread test reveals two threads are loading instead of one. The image displays the first task as fully loading one thread, while the second alternates between two cores, indicating the CPU primarily uses multipliers 51 and 52 during this test. Cinebench R23 is a genuine single-thread benchmark, loading just one thread. The screenshot shows an average multiplier of 52.74, meaning the CPU frequently switched to multiplier 53 (74% of time) and 52 (26% of time). This clearly confirms the presence and usage of the 53 multiplier.

M
mjt2789
Senior Member
483
10-29-2020, 11:54 AM
#13
This software is a tool designed for your device. I’m checking its features and will share how to obtain it. Tonight I’ll verify: 1. Reset CMOS, 2. Press F3 at startup, 3. Turn on XMP1 for RAM. I’ll post updates soon.
M
mjt2789
10-29-2020, 11:54 AM #13

This software is a tool designed for your device. I’m checking its features and will share how to obtain it. Tonight I’ll verify: 1. Reset CMOS, 2. Press F3 at startup, 3. Turn on XMP1 for RAM. I’ll post updates soon.

L
LittleMoosen
Junior Member
21
10-29-2020, 07:18 PM
#14
I gave up on stock function and full-core overclocking. I set my cache at 48, fixed the vcore to 1.335, and adjusted the turbo limits accordingly. The results show improved performance—624 single and 7550 multi scores in cpuz (5.1-5.2ghz). It still hits 5.3ghz on cinenbench single core (1420) and 5.1ghz multicore (17200). Stability tests with all-core tasks stayed consistent compared to when I used the 5.1ghz setting. It gives a bit more boost for lightly threaded jobs. I also re-enabled power-saving states like C-states and EIST, which helps it shut down when idle. According to hwinfo64, my current power usage is around 7.9w CPU package and 800MHz. Edits: names in the benchmarks were changed.
L
LittleMoosen
10-29-2020, 07:18 PM #14

I gave up on stock function and full-core overclocking. I set my cache at 48, fixed the vcore to 1.335, and adjusted the turbo limits accordingly. The results show improved performance—624 single and 7550 multi scores in cpuz (5.1-5.2ghz). It still hits 5.3ghz on cinenbench single core (1420) and 5.1ghz multicore (17200). Stability tests with all-core tasks stayed consistent compared to when I used the 5.1ghz setting. It gives a bit more boost for lightly threaded jobs. I also re-enabled power-saving states like C-states and EIST, which helps it shut down when idle. According to hwinfo64, my current power usage is around 7.9w CPU package and 800MHz. Edits: names in the benchmarks were changed.

M
mennogerben
Member
199
10-30-2020, 10:19 AM
#15
I've reinitialized the BIOS and followed these steps: 1. Pressed F3 at startup 2. Turned on XMP 1. Chose MCE with Auto Let BIOS Decide. The outcomes were: all core boosts reached 4.8GHz, which is quite modest compared to others.
M
mennogerben
10-30-2020, 10:19 AM #15

I've reinitialized the BIOS and followed these steps: 1. Pressed F3 at startup 2. Turned on XMP 1. Chose MCE with Auto Let BIOS Decide. The outcomes were: all core boosts reached 4.8GHz, which is quite modest compared to others.

A
Andk_
Junior Member
49
10-30-2020, 12:46 PM
#16
@unclewebb Here's another screenshot showing package power
A
Andk_
10-30-2020, 12:46 PM #16

@unclewebb Here's another screenshot showing package power

M
mini_man3000
Member
149
10-30-2020, 08:48 PM
#17
Started with MCE activated and XMP1 running. On the desktop, the machine powered off and I attempted a reboot using Q Code 7F. Disconnected power, restarted, and everything appeared fine now. Odd.
M
mini_man3000
10-30-2020, 08:48 PM #17

Started with MCE activated and XMP1 running. On the desktop, the machine powered off and I attempted a reboot using Q Code 7F. Disconnected power, restarted, and everything appeared fine now. Odd.

K
KochJus7in
Member
59
11-03-2020, 01:46 AM
#18
I performed MCE on my gigabyte board and it set the clock to 5.3GHz all-core and 1.56V ROUTE. I reversed it using the speed.
K
KochJus7in
11-03-2020, 01:46 AM #18

I performed MCE on my gigabyte board and it set the clock to 5.3GHz all-core and 1.56V ROUTE. I reversed it using the speed.

D
dummyiis
Member
68
11-05-2020, 03:36 PM
#19
It seems like you're expressing a concern. Could you clarify what you mean?
D
dummyiis
11-05-2020, 03:36 PM #19

It seems like you're expressing a concern. Could you clarify what you mean?

K
Kamerad_Luca
Junior Member
12
11-07-2020, 05:39 AM
#20
I think the motherboards all just go nuts doing their own thing. IMO best to manually adjust it or run at default without any enhancements. Part of me thinks its intentional to burn out your CPUs so you have to buy more motherboards and upgrade sooner.
K
Kamerad_Luca
11-07-2020, 05:39 AM #20

I think the motherboards all just go nuts doing their own thing. IMO best to manually adjust it or run at default without any enhancements. Part of me thinks its intentional to burn out your CPUs so you have to buy more motherboards and upgrade sooner.

Pages (3): Previous 1 2 3 Next