F5F Stay Refreshed Hardware Desktop Adjusting the bar size on AMD APUs

Adjusting the bar size on AMD APUs

Adjusting the bar size on AMD APUs

S
Strafeliner
Member
165
05-19-2017, 06:41 AM
#1
Checking if increasing your APU's bar size from 512MB to 4GB is safe depends on your system configuration. Your setup includes an AMD Ryzen 5 4600G and a Gigabyte B450M DS3H V2 with 16GB RAM, but the maximum VRAM is only 2GB. Since you're still saving for a new GPU, you might want to explore other performance hacks while you wait.
S
Strafeliner
05-19-2017, 06:41 AM #1

Checking if increasing your APU's bar size from 512MB to 4GB is safe depends on your system configuration. Your setup includes an AMD Ryzen 5 4600G and a Gigabyte B450M DS3H V2 with 16GB RAM, but the maximum VRAM is only 2GB. Since you're still saving for a new GPU, you might want to explore other performance hacks while you wait.

J
Jayden32805
Member
212
05-19-2017, 11:34 AM
#2
Resizable Bar enables the CPU to treat a GPU's memory as one continuous area. This concept applies best to dedicated GPUs, which possess their own VRAM. An integrated GPU already uses part of the CPU's memory. Expanding RAM for an iGPU isn't the same—it just adds more storage while reducing what the CPU has available. It’s secure but could lead to running low on RAM if you’re nearly full.
J
Jayden32805
05-19-2017, 11:34 AM #2

Resizable Bar enables the CPU to treat a GPU's memory as one continuous area. This concept applies best to dedicated GPUs, which possess their own VRAM. An integrated GPU already uses part of the CPU's memory. Expanding RAM for an iGPU isn't the same—it just adds more storage while reducing what the CPU has available. It’s secure but could lead to running low on RAM if you’re nearly full.

M
Muttbutt0219
Junior Member
5
05-19-2017, 04:47 PM
#3
It won't affect power usage significantly. The 2GB VRAM limit is usually based on hardware constraints, not just general performance. You can't really exceed it unless you change system settings or use specific software optimizations.
M
Muttbutt0219
05-19-2017, 04:47 PM #3

It won't affect power usage significantly. The 2GB VRAM limit is usually based on hardware constraints, not just general performance. You can't really exceed it unless you change system settings or use specific software optimizations.

R
Randmfrogman
Member
116
05-20-2017, 10:48 AM
#4
It shouldn't affect the desktop at all. By default, the CPU can read from GPU memory in 256 MB sections. For instance, if it needs to transfer 2 GB into VRAM, that would require 8 times 256 MB writes. When reBar is active, the CPU can read the entire GPU VRAM at once, allowing the transfer to happen in one go. This may boost performance in certain games. Greater performance usually results in higher FPS, which also means more power consumption. As mentioned earlier, reBar doesn't alter the amount of memory your iGPU actually has. If it's capped at 2 GB, enabling reBar won't change that. Improvements for iGPUs generally come from faster RAM (XMP/manual overclock). But since iGPUs are relatively weak, the gains are usually limited to around 2 GB. This is why you're often stuck with a maximum of that size. The iGPU just isn't powerful enough to benefit significantly beyond that.
R
Randmfrogman
05-20-2017, 10:48 AM #4

It shouldn't affect the desktop at all. By default, the CPU can read from GPU memory in 256 MB sections. For instance, if it needs to transfer 2 GB into VRAM, that would require 8 times 256 MB writes. When reBar is active, the CPU can read the entire GPU VRAM at once, allowing the transfer to happen in one go. This may boost performance in certain games. Greater performance usually results in higher FPS, which also means more power consumption. As mentioned earlier, reBar doesn't alter the amount of memory your iGPU actually has. If it's capped at 2 GB, enabling reBar won't change that. Improvements for iGPUs generally come from faster RAM (XMP/manual overclock). But since iGPUs are relatively weak, the gains are usually limited to around 2 GB. This is why you're often stuck with a maximum of that size. The iGPU just isn't powerful enough to benefit significantly beyond that.