• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

DirectStorage Benchmark

Joined
Nov 26, 2021
Messages
1,908 (1.43/day)
Location
Mississauga, Canada
Processor Ryzen 7 5700X
Motherboard ASUS TUF Gaming X570-PRO (WiFi 6)
Cooling Noctua NH-C14S (two fans)
Memory 2x16GB DDR4 3200
Video Card(s) Reference Vega 64
Storage Intel 665p 1TB, WD Black SN850X 2TB, Crucial MX300 1TB SATA, Samsung 830 256 GB SATA
Display(s) Nixeus NX-EDG27, and Samsung S23A700
Case Fractal Design R5
Power Supply Seasonic PRIME TITANIUM 850W
Mouse Logitech
VR HMD Oculus Rift
Software Windows 11 Pro, and Ubuntu 20.04
PC Games Hardware tested various GPUs and a 12900k using the DirectStorage benchmark provided by Compusemble. I'm wondering if any of the forum members has tried other GPUs and CPUs. I tested it on my main system with the following results:

DiskGPUCPU
SATA
1.6​
1.6​
Intel NVMe
4.6​
4.8​
WD NVMe
4.6​
4.8​

It turns out that a 5700X is faster than a Vega 64 at this, but the GPU never seemed to be reaching peak speeds. CPU utilization was pegged at 100% when the test was active while the GPU struggled to get out of sub 500 MHz clock speeds and 20 W power consumption. I would be interested in results using older GPUs and CPUs than the ones in PC Games Hardware's article.
 
PC Games Hardware tested various GPUs and a 12900k using the DirectStorage benchmark provided by Compusemble. I'm wondering if any of the forum members has tried other GPUs and CPUs. I tested it on my main system with the following results:

DiskGPUCPU
SATA
1.6​
1.6​
Intel NVMe
4.6​
4.8​
WD NVMe
4.6​
4.8​

It turns out that a 5700X is faster than a Vega 64 at this, but the GPU never seemed to be reaching peak speeds. CPU utilization was pegged at 100% when the test was active while the GPU struggled to get out of sub 500 MHz clock speeds and 20 W power consumption. I would be interested in results using older GPUs and CPUs than the ones in PC Games Hardware's article.
If CPU is faster than GPU, then this version of DS isn't working as intended. Its purpose is to off-load the texture loading to the GPU's VRAM and thus, circumventing the CPU which is slower in that task.
 
If CPU is faster than GPU, then this version of DS isn't working as intended. Its purpose is to off-load the texture loading to the GPU's VRAM and thus, circumventing the CPU which is slower in that task.

Do you happen to know of any GPU memory bandwidth and space overhead introduced by DirectStorage? If DirectStorage uses the GPU's VRAM then it must also use some memory bandwidth which could be a problem for bandwidth strapped GPUs like the 4070 Ti.
 
If this is the avocado DS 1.1 benchmark that's been going around, here's mine: Gen 3 WD Green SN350 480 GB

1674624825243.png


1674624950900.png


I also tested on the old 160 GB 320, it takes 11.5 seconds on average, but the avocados don't load.
 
Do you happen to know of any GPU memory bandwidth and space overhead introduced by DirectStorage? If DirectStorage uses the GPU's VRAM then it must also use some memory bandwidth which could be a problem for bandwidth strapped GPUs like the 4070 Ti.
No one can know that before any comparison between similar-tier GPUs being tested.
 
Back
Top