F5F Stay Refreshed Hardware Desktop Benchmarking in the lowest 5%

Benchmarking in the lowest 5%

Benchmarking in the lowest 5%

Pages (3): Previous 1 2 3 Next
C
Conk
Member
114
02-19-2026, 04:26 PM
#11
To modify virtual memory settings in Windows 11, open Settings > System > About > Advanced system settings, then go to Performance and click Settings. Under Advanced, select the Advanced tab and click Change for Virtual Memory. Disable the option that says "Automatically manage paging file size for all drives," pick your drive, set a custom size, input the initial and maximum sizes in MB, and confirm by clicking Set and OK. Restart the system for the changes to apply.
C
Conk
02-19-2026, 04:26 PM #11

To modify virtual memory settings in Windows 11, open Settings > System > About > Advanced system settings, then go to Performance and click Settings. Under Advanced, select the Advanced tab and click Change for Virtual Memory. Disable the option that says "Automatically manage paging file size for all drives," pick your drive, set a custom size, input the initial and maximum sizes in MB, and confirm by clicking Set and OK. Restart the system for the changes to apply.

U
UnaAdem
Junior Member
4
02-21-2026, 07:32 AM
#12
You're welcome, but it seems the D: Drive (980pro) is already set as the virtual memory drive. Appreciate your assistance.
U
UnaAdem
02-21-2026, 07:32 AM #12

You're welcome, but it seems the D: Drive (980pro) is already set as the virtual memory drive. Appreciate your assistance.

T
64
02-21-2026, 07:56 AM
#13
well not sure there isn't something wrong here.
Video
Says they're going at 1440p with 9gb ram usage. and 2666mhz.
If your memory usage is at 90% maybe a dimm isn't working properly. And the power supply is really bad. That could also be messin with the gpu. The chief difference in the example tube is they have a decent psu - evga supernova mg. Probably an older model but they tend to hover around tier A or B point. Vid is from 2022. Also a 3060ti rather than 3070ti.
Says his settings are on very high. maybe you've cranked them up to ultra or whatever the maximum is.
Hard to find an exact comparison. But you should be zipping along.
You might not need to upgrade the entire system - just the psu. I don't think there's a purely software fix - but you might want to clear drivers with
ddu
and reinstall the latest.
As for memory testing you'd have to look it up - I don't do that often so I don't know it off the top of my head. There's different versions of memtest86 around now or something but last time I used test mem 5 with hi 20 plugin. just don't ask me for detailed instructions. I know how to do it when it's in front of me.
A 750w quality psu is preferred for the 3070ti. As you can see the results of using a poorly rated psu are rather disappointing.
*do not mix'n'match modular cables from modular power supplies. Pull all the old ones and use only the cables supplied with a new one.
T
TheJimmyIsHere
02-21-2026, 07:56 AM #13

well not sure there isn't something wrong here.
Video
Says they're going at 1440p with 9gb ram usage. and 2666mhz.
If your memory usage is at 90% maybe a dimm isn't working properly. And the power supply is really bad. That could also be messin with the gpu. The chief difference in the example tube is they have a decent psu - evga supernova mg. Probably an older model but they tend to hover around tier A or B point. Vid is from 2022. Also a 3060ti rather than 3070ti.
Says his settings are on very high. maybe you've cranked them up to ultra or whatever the maximum is.
Hard to find an exact comparison. But you should be zipping along.
You might not need to upgrade the entire system - just the psu. I don't think there's a purely software fix - but you might want to clear drivers with
ddu
and reinstall the latest.
As for memory testing you'd have to look it up - I don't do that often so I don't know it off the top of my head. There's different versions of memtest86 around now or something but last time I used test mem 5 with hi 20 plugin. just don't ask me for detailed instructions. I know how to do it when it's in front of me.
A 750w quality psu is preferred for the 3070ti. As you can see the results of using a poorly rated psu are rather disappointing.
*do not mix'n'match modular cables from modular power supplies. Pull all the old ones and use only the cables supplied with a new one.

P
PinkaminaPie
Member
113
02-22-2026, 12:36 AM
#14
So a bit more investigation revealed that the issue wasn't just a random glitch, but a more complex problem. I applied some standard solutions and achieved impressive performance at around 60 FPS, even better with an unlocked frame rate. Checking other titles gave me the expected frame rates, confirming it was a game-related issue. Thanks a lot for your support!
P
PinkaminaPie
02-22-2026, 12:36 AM #14

So a bit more investigation revealed that the issue wasn't just a random glitch, but a more complex problem. I applied some standard solutions and achieved impressive performance at around 60 FPS, even better with an unlocked frame rate. Checking other titles gave me the expected frame rates, confirming it was a game-related issue. Thanks a lot for your support!

C
crazyant
Member
165
02-22-2026, 05:25 PM
#15
hard to find an exact comparison .
here's one with 16gb ddr4 :
no man's sky
they are using 8.5gb of ram out of 16gb for 1440p ultra. with a 3070. Also has applied vulkan as the graphics api. 110fps.
Frame rate calculator
says you should be getting around 100fps. Something is off and if you circle the wagons and even go so far as a clean os install - and the problem persists I would really suspect the psu.
C
crazyant
02-22-2026, 05:25 PM #15

hard to find an exact comparison .
here's one with 16gb ddr4 :
no man's sky
they are using 8.5gb of ram out of 16gb for 1440p ultra. with a 3070. Also has applied vulkan as the graphics api. 110fps.
Frame rate calculator
says you should be getting around 100fps. Something is off and if you circle the wagons and even go so far as a clean os install - and the problem persists I would really suspect the psu.

T
Trexter
Junior Member
10
02-23-2026, 10:11 AM
#16
Tried it with my fiancees power supply, corsair rm750 only one year old, made no difference. She also has a 32 gb ram vs my 16, also made no noticeable difference. Im sure something is wrong. I just cant pinpoint it. Even with her ram i was at 60 percent usage. So something is definitely using too much ram, but according to task manager, only chrome, desktop windows manager, spotify, and windows defender are the only things taking up any real memory usage. Event then its only about 1-2 gb at most.
T
Trexter
02-23-2026, 10:11 AM #16

Tried it with my fiancees power supply, corsair rm750 only one year old, made no difference. She also has a 32 gb ram vs my 16, also made no noticeable difference. Im sure something is wrong. I just cant pinpoint it. Even with her ram i was at 60 percent usage. So something is definitely using too much ram, but according to task manager, only chrome, desktop windows manager, spotify, and windows defender are the only things taking up any real memory usage. Event then its only about 1-2 gb at most.

M
Minermaster43
Member
60
02-24-2026, 02:57 AM
#17
well could still suspect the psu over time was dinging up the gpu. The psu notes on spl's list says it can produce 'fatal voltage drops'. Bit uh terse and non descriptive but maybe chose fatal for impact.
gpuz
- should tell you if all the lights are on your gpu vram. also has a log you can set going in the background
try
userbenchmark
that will give you more details of which component is under performing. Some people hate it since they think it's not a scientific benchmark but I think it can provide useful info. Also they think it's 'malware' though it's probably just their coding methods trigger false positives. You don't have to use it if think you're risking something.
Or
Geekbench 6
has a gpu compute and separate cpu benchmark if you loathe userbenchmark.
Also
msi afterburner & rtss
can give you an on screen overlay to monitor stats in game.
and
hwinfo
can log your power sensors to a .csv which you can load into a spreadsheet to see if you are getting voltage sags.
Lots of tools around. Probably better to try and diagnose than swapping parts around all the time. less you handle them the better in general.
But I suppose it could be quicker to simply try the gpu in another system. just fidgety - clean, install drivers, put the other gpu back when you're done, etc. etc. vs. the risk of breaking something else accidentally.
could also wonder if stwessing it with a non recommended psu is a good idea. the bench tests are shorter than a serious stress test but if the psu isn't up to it could cause a bit of mayhem.
So probably check the voltages first with hwinfo.
M
Minermaster43
02-24-2026, 02:57 AM #17

well could still suspect the psu over time was dinging up the gpu. The psu notes on spl's list says it can produce 'fatal voltage drops'. Bit uh terse and non descriptive but maybe chose fatal for impact.
gpuz
- should tell you if all the lights are on your gpu vram. also has a log you can set going in the background
try
userbenchmark
that will give you more details of which component is under performing. Some people hate it since they think it's not a scientific benchmark but I think it can provide useful info. Also they think it's 'malware' though it's probably just their coding methods trigger false positives. You don't have to use it if think you're risking something.
Or
Geekbench 6
has a gpu compute and separate cpu benchmark if you loathe userbenchmark.
Also
msi afterburner & rtss
can give you an on screen overlay to monitor stats in game.
and
hwinfo
can log your power sensors to a .csv which you can load into a spreadsheet to see if you are getting voltage sags.
Lots of tools around. Probably better to try and diagnose than swapping parts around all the time. less you handle them the better in general.
But I suppose it could be quicker to simply try the gpu in another system. just fidgety - clean, install drivers, put the other gpu back when you're done, etc. etc. vs. the risk of breaking something else accidentally.
could also wonder if stwessing it with a non recommended psu is a good idea. the bench tests are shorter than a serious stress test but if the psu isn't up to it could cause a bit of mayhem.
So probably check the voltages first with hwinfo.

H
harm2046
Member
219
02-24-2026, 07:43 PM
#18
If MSI Afterburner or another GPU tool is installed, remove it first. Then run DDU in safe mode to eliminate Nvidia and Radeon-related files, and reinstall the newest Nvidia driver (after that, you can reinstall Afterburner if necessary). Also, ensure your motherboard has the latest AMD chipset driver.

You shouldn't depend on CPU load to detect bottlenecks because CPUs rarely run at full capacity during games—being at 50% load doesn’t always mean a problem. The GPU, however, should consistently operate near 100%. A 60% utilization suggests an obstruction to achieving peak performance, which could stem from CPU limitations, FPS restrictions, driver issues, or faulty hardware. A poor power supply is unlikely to cause frame rate issues. This explanation doesn’t quite add up.
H
harm2046
02-24-2026, 07:43 PM #18

If MSI Afterburner or another GPU tool is installed, remove it first. Then run DDU in safe mode to eliminate Nvidia and Radeon-related files, and reinstall the newest Nvidia driver (after that, you can reinstall Afterburner if necessary). Also, ensure your motherboard has the latest AMD chipset driver.

You shouldn't depend on CPU load to detect bottlenecks because CPUs rarely run at full capacity during games—being at 50% load doesn’t always mean a problem. The GPU, however, should consistently operate near 100%. A 60% utilization suggests an obstruction to achieving peak performance, which could stem from CPU limitations, FPS restrictions, driver issues, or faulty hardware. A poor power supply is unlikely to cause frame rate issues. This explanation doesn’t quite add up.

I
iStrafeRunner
Member
169
02-25-2026, 12:26 PM
#19
gravity is merciless. if the gpu is a brick and the psu is a pedestrian, a brick falling from a height simply doesn't matter. 0 frames per second. Only penny-related issues lead to minor skin irritation.
I
iStrafeRunner
02-25-2026, 12:26 PM #19

gravity is merciless. if the gpu is a brick and the psu is a pedestrian, a brick falling from a height simply doesn't matter. 0 frames per second. Only penny-related issues lead to minor skin irritation.

E
Enderboss1449
Member
203
02-26-2026, 05:13 AM
#20
CPU: AMD Ryzen 5 5600X (4.75GHz Curve Optimizer)
CPU Cooler: Noctua NH-D15S
Memory: Asus X570-F Strix
RAM: 2x8Gb Patriot Viper 4400Mhz CL19 (3800 MHz, CL16 / Tightened Sub-timings)
GPU: AMD RADEON RX 6800 @2400/2100 MHz
HDD: 1x2TB + 1x4TB
SSD: 240Gb Kingston Nvme + 500Gb Toshiba Nvme
Power Supply: Corsair RM750X (White) 750W 80+ Gold
Case: Thermaltake View 51 TG
Performance: 110fps guy – actually a well-optimized setup for this machine. Likely around 50fps (roughly) for a 3070ti.
E
Enderboss1449
02-26-2026, 05:13 AM #20

CPU: AMD Ryzen 5 5600X (4.75GHz Curve Optimizer)
CPU Cooler: Noctua NH-D15S
Memory: Asus X570-F Strix
RAM: 2x8Gb Patriot Viper 4400Mhz CL19 (3800 MHz, CL16 / Tightened Sub-timings)
GPU: AMD RADEON RX 6800 @2400/2100 MHz
HDD: 1x2TB + 1x4TB
SSD: 240Gb Kingston Nvme + 500Gb Toshiba Nvme
Power Supply: Corsair RM750X (White) 750W 80+ Gold
Case: Thermaltake View 51 TG
Performance: 110fps guy – actually a well-optimized setup for this machine. Likely around 50fps (roughly) for a 3070ti.

Pages (3): Previous 1 2 3 Next