The problem with the video card (sort of like)
During the game (let's say a valorant), the stable delay to GPU 10-15 is not more, but sometimes it happens that the delay rises to 30 and the FPS drops by 30 seconds, from 200 to 50-30 and lags terribly.(Video card 1050ti, percent FX8320E, 12 GB of RAM)
Eoliy
Eoliy wrote:
percent FX8320E
This is the likely reason. Percent is sensitive to memory bandwidth. Even if the RAM works at the standard frequencies for the processor (1866), then there will be friezes, and if it is even lower (1600/1333 or lower), then it’s quite a problem + you have to look at the delays. How much are memory sticks and which ones?
Len4ik00N
By the way, is it possible to fix it somehow?
Because it started around the autumn of that year, he worked for half a year (played quite demanding games), and everything was normal.
Eoliy
One of the possible options, except for some software, is overheating of the chipset, vrm and / or the processor itself.
Your memory, of course, probably works at low speeds, but this is not a reason to drop fps from 200 to 30.
I bet on overheating, especially considering the current heat.
What motherboard do you have? Did you overclock the percent? what voltage did you put on it? What kind of cooling is installed on the process and how many fans are installed directly in the case?
If there is a thermocouple / pyrometer, measure the temperature on the BPM (it is possible from the front and back sides of the motherboard), as well as with the chipset. You can just touch it with your finger, but here at your own peril and risk, so as not to close anything, you can touch the heatsink on the chipset and vrm. Ofc, not in a simple way, but in some kind of load.
Eoliy
Eoliy wrote:
By the way, is it possible to fix it somehow?
You can - by overclocking, but here a lot depends on the board and memory strips + you need to make sure that there are no problems with heating the iron (the person wrote about this above)
close one
I also thought about overheating, because at first everything was normal, and then over time it already started to warm up more.
The radiator is very hot, (I often changed the thermal paste, due to this problem, while 1-3 days will pass after changing the thermal paste, it starts to warm up again.
(Prots did not overclock)
(There are no fans, because the block cover is open)
Eoliy
Check the CPU clock speed during such severe FPS drawdowns, MSI afterburner will help you.
I bet monitoring will show 1400 MHz.
Also, an open case is not a panacea, the iron needs flowing air for more efficient cooling. Iron heats the air around it and if this heated air is not blown out, then accordingly this heated air will heat the iron even more. And this applies not only to the processor, on your motherboard, the BPM and the chipset are also susceptible to overheating.
I also forgot to ask what kind of power supply model? If the power supply unit is normal and you are sure of it, then just reduce the pressure by a percentage. And organize a normal cooling for your gland, this also applies to the vrm zone, and not just the core.
Eoliy
BP, konesh, shit, but 1050 + fufik should pull.
Yes, it's okay, you didn’t overclock the percentage and didn’t fix the frequency at the same value, so the energy saving devices in idle time and lower its frequency. All modern and relatively modern processes work this way. Video cards too.
close one
Regarding the frequencies, I will answer tomorrow because at night (the room temperature has already fallen asleep, compared to the day), then this problem is not
Eoliy
Eoliy wrote:
During the game ... sometimes it happens that the delay rises to 30 and the FPS drops by 30 seconds, from 200 to 50-30 and lags terribly ...
... the room temperature has already fallen asleep compared to the day ), then this problem is not ...
Perhaps overheating and, as a result, throttling. Was the load on the system (MSI Afterburner, etc.) monitored during the game - temperatures, frequencies, etc.?
close one I
watched the whole movement in MSI, in general, the video card simply drops its load for 30 seconds, and freezes occur, and at the same time, when the load on the GPU drops, everything goes to the CPU. There were no special temperature rises (40-50% and percent and vidyuha).
Also, during the game, the load on the video card is 30-40% is this normal? (I just got used to when I have about 70.
Eoliy
In a normal state, in the vast majority of games, without FPS lock and other vertical syncs, the load on the video card should be 99-100% most of the time. If this is not the case, then the emphasis on performance goes to the processor / memory.
If the load of the video card sags even more, then there is an even greater emphasis in percent / memory.
What is the frequency on the CPU in times of severe drawdowns? Not when there, for example, from 90 to 60 it sags, but as you wrote above from 200 to 30 with strong lags.
close one
Is it possible to somehow change the emphasis from the processor to the video card?
Eoliy
Buy a more powerful processor and faster memory for it.
If you want to understand how it works approximately and have 10 extra minutes.
If you are interested in more details:
close one
I went into another game, which is more demanding (sniper gv contracts 2), and everything is much more stable there, 70 degrees vidyuha, 44 degrees percent, and the load on it is 40-50, and on vidyuha 99. (At the same time , there is no problem with the load drop)
BUT why is it wrong in VALORANT?
3500mhz memory idle