TechPowerUp Releases GPU-Z v2.60.0 (2024)

TechPowerUp has released version 2.60.0 of GPU-Z, a popular graphics sub-system information, monitoring, and diagnostic utility. This latest update brings significant enhancements, including full support for the Arm64 architecture and Qualcomm Snapdragon X Elite GPUs. The release also adds support for AMD Zen 5 CPU temperature monitoring and a wide range of new GPUs from NVIDIA, AMD, and Intel. Notable additions include the NVIDIA 4070 Ti Super (AD102), RTX 4070 (AD103), RTX 4060 Ti (AD104), RTX 4060 (AD106), as well as AMD Zen 5 (Strix Point and Granite Ridge), and Intel Raptor Lake U SKUs and Meteor Lake Intel Arc Graphics.

In addition to expanded hardware support, GPU-Z 2.60.0 addresses several important issues. The update fixes NVIDIA driver version reporting for some pre-2015 versions, resolves an installer problem that prevented closing running instances of GPU-Z, and corrects the "0 MHz" memory clock display on certain AMD RDNA GPUs without overclocking support. Other improvements include a small handle leak fix, added support for the Monster Notebook subvendor ID, and compatibility with new VMWare virtual GPU IDs. The installer now requires Windows 7 or newer, with appropriate messaging for unsupported systems. Users can download the latest version of TechPowerUp GPU-Z from the official TechPowerUp website to access these new features and improvements.

DOWNLOAD: TechPowerUp GPU-Z 2.60.0

Below, you can check out the complete changelog list.

• Added full support for Arm64 and Qualcomm Snapdragon X Elite and other Arm64 GPUs
• Added support for AMD Zen 5 CPU temperature monitoring
• Fixed NVIDIA driver version reporting for some specific versions, due to leading zeros (before 2015)
• Fixed installer isn't able to close running instances of GPU-Z
• Installer now requires Windows 7 and newer, appropriate messaging is displayed when not supported
• Fixed small handle leak
• Fixed "0 MHz" memory clock display on some AMD RDNA GPUs without overclocking support
• Added Monster Notebook subvendor Id
• Added support for NVIDIA RTX 4070 Ti Super (AD102), RTX 4070 (AD103), RTX 4060 Ti (AD104), RTX 4060 (AD106), A1000, A400, RTX 500 Ada Laptop, RTX A2000 Ada Embedded, Drive PG199, H100 NVL
• Added support for AMD Zen 5 (Strix Point and Granite Ridge), Phoenix Radeon 740M
• Added support for Intel Raptor Lake U SKUs, Meteor Lake Intel Arc Graphics
• Added support for new VMWare virtual GPU Id

View at TechPowerUp Main Site

Slight problem. File isn't available:
@W1zzard

windwhirl said:

Slight problem. File isn't available:
@W1zzard

View attachment 359228

Click to expand...

Uploading ASAP!!!

My 2.57.0 works with no flaw so perhaps I'll update another time.

Whoever is behind this tool: thank you. Nice software.

Always nice seeing a new version out

As a side note, are there any plans for a linux release now that full arm64 support has been added? I'd love to get this on some of my ARM-based SBCs running Debian to see the guts of the integrated GPUs.

windwhirl said:

Slight problem. File isn't available:

Click to expand...

Fixed

Tropick said:

As a side note, are there any plans for a linux release now that full arm64 support has been added? I'd love to get this on some of my ARM-based SBCs running Debian to see the guts of the integrated GPUs.

Click to expand...

Absolutely no plans for a Linux version, but I have plans for a big UI refresh

W1zzard said:

Fixed

Absolutely no plans for a Linux version, but I have plans for a big UI refresh

Click to expand...

Thanks for the update and the information.

Sweet! ARM support. Notice the Adreno 690 vs 680 in one of the tabs. Thats a Qualcomm issue. Its listed like this everywhere in Windows, dxdiag reports, etc.
I like the "Not using PCI-Express" notice in REBAR.

Validation: https://www.techpowerup.com/gpuz/details/ubd9r

fec32a4de said:

Sweet! ARM support. Notice the Adreno 690 vs 680 in one of the tabs. Thats a Qualcomm issue. Its listed like this everywhere in Windows, dxdiag reports, etc.

Click to expand...

Looking good. Unfortunately Qualcomm isn't willing to reveal the ROP/TMU count for this processor. I spent like a week debating back and forth with them. They did share the number for the X Elite though, which give me hope that they will learn and release this info in all future releases

Added support for new VMWare virtual GPU Id

Click to expand...

Since I've been using GPU-Z for a long time this change inspired me to contribute. At the moment GPU-Z doesn't support the graphical options of Oracle Virtualbox nor QEMU, which when used with Linux's KVM is probably the most popular virtualization toolset for that OS.

Oracle Virtualbox 7.0.20 with a Windows 10 guest OS with latest guest additions has 3 graphical options, but only the first one is officially supported and allows for limited 3D acceleration:

1) PCI\VEN_80EE&DEV_BEEF&SUBSYS_040515AD&REV_00 - Graphics Controller: VBoxSVGA, Device Name: "VirtualBox Graphics Adapter (WDDM)", enabling 3D acceleration doesn't change this
2) PCI\VEN_80EE&DEV_BEEF&SUBSYS_00000000&REV_00 - Graphics Controller: VBoxVGA, Device Name: "VirtualBox Graphics Adapter (WDDM)", 3D acceleration disabled, enabling it changes to VBoxSVGA
3) PCI\VEN_15AD&DEV_0405&SUBSYS_040515AD&REV_00 - Graphics Controller: VMSVGA, Device Name: "Microsoft Basic Display Adapter" - I didn't install VMware Tools, 3D acceleration disabled, enabling it changes to VBoxSVGA

1) and 2) show empty in GPU-Z while 3) shows up as VMware which is what it was designed to do.

QEMU 9.0 under Proxmox VE 8.2.4 which exposes the most popular of QEMU graphics adapters (with the latest VirtIO driver ISO installed), Windows 10 as well:

1) PCI\VEN_1234&DEV_1111&SUBSYS_11001AF4&REV_02 - std - Standard VGA, Device Name: "Microsoft Basic Display Adapter" even with virtio tools
2) PCI\VEN_15AD&DEV_0405&SUBSYS_040515AD&REV_00 - vmware - VMware compatible, Device Name: "Microsoft Basic Display Adapter" - I didn't install VMware Tools
3) PCI\VEN_1B36&DEV_0100&SUBSYS_11001AF4&REV_05 - qxl - SPICE, Device Name: "Red Hat QXL controller"
4) PCI\VEN_1B36&DEV_0100&SUBSYS_11001AF4&REV_05 - qxl2 - SPICE dual monitor, Device Name: "Red Hat QXL controller", adds 2 controllers
5) PCI\VEN_1B36&DEV_0100&SUBSYS_11001AF4&REV_05 - qxl3 - SPICE three monitors, Device Name: "Red Hat QXL controller", adds 3 controllers
6) PCI\VEN_1B36&DEV_0100&SUBSYS_11001AF4&REV_05 - qxl4 - SPICE four monitors, Device Name: "Red Hat QXL controller", adds 4 controllers
7) PCI\VEN_1AF4&DEV_1050&SUBSYS_11001AF4&REV_01 - virtio - VirtIO-GPU, Device Name: "Red Hat VirtIO GPU DOD controller"
8) PCI\VEN_1AF4&DEV_1050&SUBSYS_11001AF4&REV_01 - virtio-gl - VirGL GPU, Device Name: "Red Hat VirtIO GPU DOD controller"

Ad 8) This should allow for limited OpenGL support from the host GPU, but I'm not sure that support is even available under Windows.
Just like with Virtualbox only 2) shows up in GPU-Z.

There's also Xen, but I don't have any infrastructure in place to test what they provide. From the driver packages it doesn't look like they have anything special for graphics.

Beginner Macro Device said:

My 2.57.0 works with no flaw so perhaps I'll update another time.

Click to expand...

It was fun installing the RTX 3050 6GB for the first time as GPU-Z 2.58.0 was giving me a look, asking WTF I was doing with unknown hardware. 2.59.0 fixed that up.

Lew Zealand said:

It was fun installing the RTX 3050 6GB for the first time as GPU-Z 2.58.0 was giving me a look, asking WTF I was doing with unknown hardware. 2.59.0 fixed that up.

Click to expand...

Yeah I know but my future GPU does not yet exist. Don't see any sense in upgrading a 6700 XT to something available today. Even 4090 doesn't sound too attractive. Especially because I'm not really rich.

W1zzard said:

Absolutely no plans for a Linux version, but I have plans for a big UI refresh

Click to expand...

Let's hope you will make it beautiful and modern. Maybe include Aero look?

BTW, there is a need to rearrange the information on the main Graphics Card tab. Why not replace the AMD/nvidia brand logo with a photo of the actual graphics card in use?
Also, Technologies must be somewhere up, not at the bottom.
You can order it this way:
1 row. Name
2 row. GPU; Die Size; Transistors (count in Bln, Mln)
3 row. Manufacturing Process; Revision; Release Date
4 row. AIB; Device ID
5 row. VBIOS Version
6 row. Shaders; Shaders Clock; ROPs/TMUs Quantity
7 row. VRAM Type; Bus Width; Memory Clock
7 row. Bus Interface;
8 row. Pixel & Texture Fillrate; Memory Bandwidth
9 row. DirectX Support; Technologies; Computing
10 row. Driver Data

Beginner Macro Device said:

Yeah I know but my future GPU does not yet exist. Don't see any sense in upgrading a 6700 XT to something available today. Even 4090 doesn't sound too attractive. Especially because I'm not really rich.

Click to expand...

There will be a very long wait. I would recommend wait for a 2nm AMD Radeon in the ballpark of Radeon RX 7900 XT 20 GB market tier.

One change that was not made in 2.60.0 was the speed ordering:

GPU Clock/Memory/Boost

• The first refers to the GPU
• The second to the memory
• The third to the GPU again

fec32a4de said:

Sweet! ARM support. Notice the Adreno 690 vs 680 in one of the tabs. Thats a Qualcomm issue. Its listed like this everywhere in Windows, dxdiag reports, etc.
I like the "Not using PCI-Express" notice in REBAR.

Validation: https://www.techpowerup.com/gpuz/details/ubd9r

View attachment 359238View attachment 359234View attachment 359233

View attachment 359235View attachment 359236

Click to expand...

And that there's no hardware acceleration, like it's the pre-Pentium II days, LOL.

W1zzard said:

Looking good. Unfortunately Qualcomm isn't willing to reveal the ROP/TMU count for this processor. I spent like a week debating back and forth with them. They did share the number for the X Elite though, which give me hope that they will learn and release this info in all future releases

Click to expand...

GPU-Z for ARM64 works extremely well. The fact that Qualcоmm keeps unnecessary secrets and/or tends to obfuscate info is on them. Even M$ doesn't know what do to here.

If you need any logs or info u can contact me any time.

But for now, have the lower spec 7c, also from Samsung. Sam didn't want to sing, but in the end Samsung - (i know its a different version of the joke, pls dont kill me).

Validation: https://www.techpowerup.com/gpuz/details/cgs7g

ARF said:

There will be a very long wait. I would recommend wait for a 2nm AMD Radeon in the ballpark of Radeon RX 7900 XT 20 GB market tier.

Click to expand...

Who said I'm in a hurry? I can wait almost forever, I got plenty o'business to catch up so another four or six years of not buying a GPU won't be a problem. AMD, NV, Intel... I will buy whatever I feel most feature-rich, even if lacking in strict perf per $. I just wanna be able to tick EVERY box and see if it's good rather than having a half boxes unavailable and a quarter, working poorly. That's why I'd buy a 4080 over 7900 XTX today because the latter can't do RT as smoothly, can't do DLSS whatsoever, and is a little worse in energy efficiency (however, not that I care about that lmao) despite being a smidge better in raw performance per dollar.

So, yeah, before that, I'll only update GPU-Z if I delete it by accident or stumble upon new cool features.

v12dock said:

Any plans on releasing a native ARM64 build of GPU-Z?

Click to expand...

No plans at this time, the current solution works well enough and there's no need to confuse people with 3 different downloads.

Shrek said:

One change that was not made in 2.60.0 was the speed ordering:

Click to expand...

This will not happen until the UI rewrite, probably not even in 2024

W1zzard said:

Looking good. Unfortunately Qualcomm isn't willing to reveal the ROP/TMU count for this processor. I spent like a week debating back and forth with them. They did share the number for the X Elite though, which give me hope that they will learn and release this info in all future releases

Click to expand...

Why would they hold back such trivial information? I can't imagine.

AusWolf said:

Why would they hold back such trivial information? I can't imagine.

Click to expand...

Sometimes, stubborn pendejos are hired. Sometimes, it's just a misinterpretation of the "classified information" term: they might think they are not allowed to share this data.