Benchmarking Ubuntu vs Windows on the same HP ZBook
Laptop specs π
I have an HP ZBook 15 G3 with HP 80D5 motherboard.
CPU Information (processor):
| item | info |
|---|---|
| Name | Intel Core i7-6820HQ |
| Topology | 1 Processor, 4 Cores, 8 Threads |
| Identifier | GenuineIntel Family 6 Model 94 Stepping 3 |
| Base Frequency | 2.71 GHz |
| Cluster 1 | 4 Cores |
| Maximum Frequency | 3590 MHz |
| Package | Socket 1440 FCBGA |
| Codename | Skylake |
| L1 Instruction Cache | 32.0 KB x 4 |
| L1 Data Cache | 32.0 KB x 4 |
| L2 Cache | 256 KB x 4 |
| L3 Cache | 8.00 MB x 1 |
| Instruction Sets | sse2 sse3 pclmul fma3 sse41 aesni avx avx2 |
Memory Information (RAM):
| item | info |
|---|---|
| Size | 16.00 GB |
| Transfer Rate | 2126 MT/s |
| Type | DDR4 SDRAM |
| Channels | 1 |
Storage:
| storage | usage |
|---|---|
| 1 TB SSD | Windows 10 + a partition for common data |
| 256 GB SSD | Ubuntu Linux |
Operating systems π
I have a dualboot setup of Ubuntu 24.04 LTS (Linux distro) and Microsoft Windows 10 Pro.
I have Microsoft Windows 10 Pro (64-bit) installed on an SSD and battery power plan set to balanced/performance.
And I have Ubuntu 24.04.3 LTS installed on another SSD storage, and power plan is set to performance.
The laptop computer was connected to the charger all the time while testing and benchmarking. It was plugged in for better performance.
Web benchmark: Speedometer 3.1 π
I used the famous Speedometer 3.1 benchmark for the web technologies on Google Chrome - the #1 web browser on Windows and Linux.
Windows 10 π
The first iteration:
The second iteration:
The third iteration:
Ubuntu π
The first iteration:
The second iteration:
The third iteration:
Crunching web benchmark numbers π
Average web benchmark on Microsoft Windows 10 pro is (11.2+11.6+11.7)/3 = 11.5
Average web benchmark on Ubuntu 24.04 LTS is (11.3+11.5+11.6)/3 = 11.4666667
The web benchmark slightly better on Windows because Google (creator of Chrome) historically prioritizes Windows-specifc GPU acceleration and compiler optimizations.
Geekbench 6 π
Geekbench 6 measures a device’s CPU and GPU performance. It is the famous benchmark everyone uses. Let’s see the results and compare!
Microsoft Windows 10 pro π
| benchmark | results linking to Geekbench result page |
|---|---|
| CPU | Single-Core Score: 1216 Multi-Core Score: 3402 |
| OpenCL on Nvidia Quadro M1200 | OpenCL Score: 11707 |
| OpenCL on Intel HD Graphics 530 | OpenCL Score: 4698 |
| Vulkan on Quadro M1200 | Vulkan Score: 10346 |
| Vulkan on Intel HD Graphics 530 | Vulkan Score: 4823 |
Ubuntu 24.04.3 LTS (Canonical’s Linux distro) π
I benchmarked OpenCL on Quadro M1200 using this command:
./geekbench6 --gpu OpenCL --gpu-platform-id 0 --gpu-device-id 0
and the output was:
Geekbench 6.5.0 : https://www.geekbench.com/
Geekbench 6 requires an active internet connection and automatically uploads
benchmark results to the Geekbench Browser.
Upgrade to Geekbench 6 Pro to enable offline use and unlock other features:
https://store.primatelabs.com/v6
Enter your Geekbench 6 Pro license using the following command line:
./geekbench6 --unlock <email> <key>
[0210/121812:FATAL:nvml_library.cpp(153)] Failed to load nvmlInit_v2: /home/ubuntu/Downloads/Geekbench-6.5.0-Linux/geekbench_avx2: undefined symbol: nvmlInit_v2
System Information
Operating System Ubuntu 24.04.3 LTS
Kernel Linux 6.17.0-14-generic x86_64
Model HP HP ZBook 15 G3
Motherboard HP 80D5
BIOS HP N81 Ver. 01.62
CPU Information
Name Intel Core i7-6820HQ
Topology 1 Processor, 4 Cores, 8 Threads
Identifier GenuineIntel Family 6 Model 94 Stepping 3
Base Frequency 3.60 GHz
L1 Instruction Cache 32.0 KB x 4
L1 Data Cache 32.0 KB x 4
L2 Cache 256 KB x 4
L3 Cache 8.00 MB
Instruction Sets sse2 sse3 pclmul fma3 sse41 aesni avx avx2
Memory Information
Size 15.5 GB
OpenCL Information
Platform Vendor NVIDIA Corporation
Platform Name NVIDIA CUDA
Device Vendor NVIDIA Corporation
Device Name Quadro M1200
Device Driver Version 580.126 .16
Maximum Frequency 1148 MHz
Compute Units 5
Device Memory 3.94 GB
OpenCL
Running Background Blur
Running Face Detection
Running Horizon Detection
Running Edge Detection
Running Gaussian Blur
Running Feature Matching
Running Stereo Matching
Running Particle Physics
Uploading results to the Geekbench Browser. This could take a minute or two
depending on the speed of your internet connection.
Upload succeeded. Visit the following link and view your results online:
https://browser.geekbench.com/v6/compute/5769927
Visit the following link and add this result to your profile:
https://browser.geekbench.com/v6/compute/5769927/claim?key=870587
And I benchmarked Vulkan on Intel(R) HD Graphics 530 (SKL GT2) using this command:
./geekbench6 --gpu Vulkan --gpu-platform-id 0 --gpu-device-id 0
And the output was like this:
Geekbench 6.5.0 : https://www.geekbench.com/
Geekbench 6 requires an active internet connection and automatically uploads
benchmark results to the Geekbench Browser.
Upgrade to Geekbench 6 Pro to enable offline use and unlock other features:
https://store.primatelabs.com/v6
Enter your Geekbench 6 Pro license using the following command line:
./geekbench6 --unlock <email> <key>
System Information
Operating System Ubuntu 24.04.3 LTS
Kernel Linux 6.17.0-14-generic x86_64
Model HP HP ZBook 15 G3
Motherboard HP 80D5
BIOS HP N81 Ver. 01.62
CPU Information
Name Intel Core i7-6820HQ
Topology 1 Processor, 4 Cores, 8 Threads
Identifier GenuineIntel Family 6 Model 94 Stepping 3
Base Frequency 3.60 GHz
L1 Instruction Cache 32.0 KB x 4
L1 Data Cache 32.0 KB x 4
L2 Cache 256 KB x 4
L3 Cache 8.00 MB
Instruction Sets sse2 sse3 pclmul fma3 sse41 aesni avx avx2
Memory Information
Size 15.5 GB
Vulkan Information
Device Name Intel(R) HD Graphics 530 (SKL GT2)
Vulkan
Running Background Blur
Running Face Detection
Running Horizon Detection
Running Edge Detection
Running Gaussian Blur
Running Feature Matching
Running Stereo Matching
Running Particle Physics
Uploading results to the Geekbench Browser. This could take a minute or two
depending on the speed of your internet connection.
Upload succeeded. Visit the following link and view your results online:
https://browser.geekbench.com/v6/compute/5769955
Visit the following link and add this result to your profile:
https://browser.geekbench.com/v6/compute/5769955/claim?key=698960
And I benchmarked Vulkan on Quadro M1200 using this command:
./geekbench6 --gpu Vulkan --gpu-platform-id 0 --gpu-device-id 1
The output was like this:
Geekbench 6.5.0 : https://www.geekbench.com/
Geekbench 6 requires an active internet connection and automatically uploads
benchmark results to the Geekbench Browser.
Upgrade to Geekbench 6 Pro to enable offline use and unlock other features:
https://store.primatelabs.com/v6
Enter your Geekbench 6 Pro license using the following command line:
./geekbench6 --unlock <email> <key>
System Information
Operating System Ubuntu 24.04.3 LTS
Kernel Linux 6.17.0-14-generic x86_64
Model HP HP ZBook 15 G3
Motherboard HP 80D5
BIOS HP N81 Ver. 01.62
CPU Information
Name Intel Core i7-6820HQ
Topology 1 Processor, 4 Cores, 8 Threads
Identifier GenuineIntel Family 6 Model 94 Stepping 3
Base Frequency 3.60 GHz
L1 Instruction Cache 32.0 KB x 4
L1 Data Cache 32.0 KB x 4
L2 Cache 256 KB x 4
L3 Cache 8.00 MB
Instruction Sets sse2 sse3 pclmul fma3 sse41 aesni avx avx2
Memory Information
Size 15.5 GB
Vulkan Information
Device Name Quadro M1200
Vulkan
Running Background Blur
Running Face Detection
Running Horizon Detection
Running Edge Detection
Running Gaussian Blur
Running Feature Matching
Running Stereo Matching
Running Particle Physics
Uploading results to the Geekbench Browser. This could take a minute or two
depending on the speed of your internet connection.
Upload succeeded. Visit the following link and view your results online:
https://browser.geekbench.com/v6/compute/5770002
Visit the following link and add this result to your profile:
https://browser.geekbench.com/v6/compute/5770002/claim?key=992108
And I benchmarked Vulkan on llvmpipe (LLVM 20.1.2, 256 bits) using this command:
./geekbench6 --gpu Vulkan --gpu-platform-id 0 --gpu-device-id 2 #
and its output was:
Geekbench 6.5.0 : https://www.geekbench.com/
Geekbench 6 requires an active internet connection and automatically uploads
benchmark results to the Geekbench Browser.
Upgrade to Geekbench 6 Pro to enable offline use and unlock other features:
https://store.primatelabs.com/v6
Enter your Geekbench 6 Pro license using the following command line:
./geekbench6 --unlock <email> <key>
System Information
Operating System Ubuntu 24.04.3 LTS
Kernel Linux 6.17.0-14-generic x86_64
Model HP HP ZBook 15 G3
Motherboard HP 80D5
BIOS HP N81 Ver. 01.62
CPU Information
Name Intel Core i7-6820HQ
Topology 1 Processor, 4 Cores, 8 Threads
Identifier GenuineIntel Family 6 Model 94 Stepping 3
Base Frequency 3.60 GHz
L1 Instruction Cache 32.0 KB x 4
L1 Data Cache 32.0 KB x 4
L2 Cache 256 KB x 4
L3 Cache 8.00 MB
Instruction Sets sse2 sse3 pclmul fma3 sse41 aesni avx avx2
Memory Information
Size 15.5 GB
Vulkan Information
Device Name llvmpipe (LLVM 20.1.2, 256 bits)
Vulkan
Running Background Blur
Running Face Detection
Running Horizon Detection
Running Edge Detection
Running Gaussian Blur
Running Feature Matching
Running Stereo Matching
Running Particle Physics
Uploading results to the Geekbench Browser. This could take a minute or two
depending on the speed of your internet connection.
Upload succeeded. Visit the following link and view your results online:
https://browser.geekbench.com/v6/compute/5770074
Visit the following link and add this result to your profile:
https://browser.geekbench.com/v6/compute/5770074/claim?key=337368
| benchmark | results linking to benchmark result page |
|---|---|
| CPU | Single-Core Score: 1237 Multi-Core Score: 3669 |
| OpenCL on NVidia Quadro M1200 | OpenCL Score: 11870 |
| Vulkan on Intel(R) HD Graphics 530 (SKL GT2) | Vulkan Score: 4429 |
| Vulkan on Nvidia Quadro M1200 | Vulkan Score: 10178 |
| Vulkan on llvmpipe (LLVM 20.1.2, 256 bits) | Vulkan Score: 690 |
Crunching Geekbench numbers π
| benchmark | Windows 10 | Ubuntu 24.04 LTS |
|---|---|---|
| CPU | Single-Core Score: 1216 Multi-Core Score: 3402 | Single-Core Score: 1237 Multi-Core Score: 3669 |
| OpenCL on Nvidia Quadro M1200 | OpenCL Score: 11707 | OpenCL Score: 11870 |
| OpenCL on Intel HD Graphics 530 | OpenCL Score: 4698 | - |
| Vulkan on Nvidia Quadro M1200 | Vulkan Score: 10346 | Vulkan Score: 10178 |
| Vulkan on Intel HD Graphics 530 | Vulkan Score: 4823 | (SKL GT2) Vulkan Score: 4429 |
| Vulkan on llvmpipe (LLVM 20.1.2, 256 bits) | - | Vulkan Score: 690 |
CPU π
CPU performance on Ubuntu Linux is better than its performance on Windows 10 by ~1.7% in single core score; (1237-1216)/1216 = 0.0172697368.
CPU multicore performance is better on Ubuntu and worse on Windows 10 by ~7.8% ; (3669-3402)/3402 = 0.0784832451.
So, Linux gives you better raw performance compared to Windows 10 on the same old hardware laptop.
GPU π
Benchmarking OpenCL on Nvidia Quadro M1200 on Ubuntu vs. Windows, told use that Ubuntu’s performance is better by ~1.4% ; (11870-11707)/11707 = 0.0139232938.
But benchmarking Vulkan on Nvidia Quadro M1200 on Ubuntu vs. Windows 10, told us that performance of Microsoft Windows 10 is better by ~1.7% ; (10346-10178)/10178 = 0.0165061898. It is likely thanks to optimized proprietary drivers developed and published by NVIDIA.
And benchmarking Vulkan on Intel HD Graphics 530 on Ubuntu Linux vs. Microsoft Windows told us that Windows 10 outperform Ubuntu by ~8.9% ; (4823-4429)/4429 = 0.0889591330.
So, I conclude that OpenCL is optimized better on Linux distros such as Ubuntu 24.04.3 LTS. On the other hand, Vulkan is optimized better for Microsoft Windows.
Hashcat π
Hashcat is a tool to crack passwords using GPU acceleration. I will use Hashcat as a benchmark of how the laptop performs.
Benchmarking Hashcat on Ubuntu 24.04 LTS π
I run hashcat -I to show all supported devices:
$ hashcat -I
hashcat (v7.1.2-382-g2d71af371) starting in backend information mode
nvmlDeviceGetFanSpeed(): Not Supported
CUDA Info:
==========
CUDA.Version.: 13.0
Backend Device ID #01 (Alias: #02)
Name...........: Quadro M1200
Processor(s)...: 5
Preferred.Thrd.: 32
Clock..........: 1148
Memory.Total...: 4035 MB
Memory.Free....: 3997 MB
Memory.Unified.: 0
Local.Memory...: 48 KB
PCI.Addr.BDFe..: 0000:01:00.0
OpenCL Info:
============
OpenCL Platform ID #1
Vendor..: NVIDIA Corporation
Name....: NVIDIA CUDA
Version.: OpenCL 3.0 CUDA 13.0.97
Backend Device ID #02 (Alias: #01)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: Quadro M1200
Version........: OpenCL 3.0 CUDA
Processor(s)...: 5
Preferred.Thrd.: 32
Clock..........: 1148
Memory.Total...: 4035 MB (limited to 1008 MB allocatable in one block)
Memory.Free....: 3997 MB
Memory.Unified.: 0
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 580.126.16
PCI.Addr.BDF...: 01:00.0
OpenCL Platform ID #2
Vendor..: Intel(R) Corporation
Name....: Intel(R) OpenCL Graphics
Version.: OpenCL 3.0
Backend Device ID #03
Type...........: GPU
Vendor.ID......: 8
Vendor.........: Intel(R) Corporation
Name...........: Intel(R) HD Graphics 530
Version........: OpenCL 3.0 NEO
Processor(s)...: 8
Preferred.Thrd.: 7
Clock..........: 1050
Memory.Total...: 14468 MB (limited to 2047 MB allocatable in one block)
Memory.Free....: 7234 MB
Memory.Unified.: 1
Local.Memory...: 64 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 23.43.027642
OpenCL Platform ID #3
Vendor..: The pocl project
Name....: Portable Computing Language
Version.: OpenCL 3.0 PoCL 5.0+debian Linux, None+Asserts, RELOC, SPIR, LLVM 16.0.6, SLEEF, DISTRO, POCL_DEBUG
Backend Device ID #04
Type...........: CPU
Vendor.ID......: 128
Vendor.........: GenuineIntel
Name...........: cpu-haswell-Intel(R) Core(TM) i7-6820HQ CPU @ 2.70GHz
Version........: OpenCL 3.0 PoCL HSTR: cpu-x86_64-pc-linux-gnu-haswell
Processor(s)...: 8
Preferred.Thrd.: 1
Clock..........: 3600
Memory.Total...: 13822 MB (limited to 2048 MB allocatable in one block)
Memory.Free....: 6911 MB
Memory.Unified.: 1
Local.Memory...: 256 KB
OpenCL.Version.: OpenCL C 1.2 PoCL
Driver.Version.: 5.0+debian
Run the Hashcat benchmark with hashcat -b -m 22000 --backend-ignore-cuda like this:
$ hashcat -b -m 22000 --backend-ignore-cuda
hashcat (v7.1.2-382-g2d71af371) starting in benchmark mode
Benchmarking uses hand-optimized kernel code by default.
You can use it in your cracking session by setting the -O option.
Note: Using optimized kernel code limits the maximum supported password length.
To disable the optimized kernel code in benchmark mode, use the -w option.
Kernel /usr/local/share/hashcat/OpenCL/m22000-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel
nvmlDeviceGetFanSpeed(): Not Supported
OpenCL API (OpenCL 3.0 CUDA 13.0.97) - Platform #1 [NVIDIA Corporation]
=======================================================================
* Device #01: Quadro M1200, 4035/4035 MB (1008 MB allocatable), 5MCU
OpenCL API (OpenCL 3.0 ) - Platform #2 [Intel(R) Corporation]
=============================================================
* Device #02: Intel(R) HD Graphics 530, 7234/14468 MB (2047 MB allocatable), 8MCU
OpenCL API (OpenCL 3.0 PoCL 5.0+debian Linux, None+Asserts, RELOC, SPIR, LLVM 16.0.6, SLEEF, DISTRO, POCL_DEBUG) - Platform #3 [The pocl project]
==================================================================================================================================================
* Device #03: cpu-haswell-Intel(R) Core(TM) i7-6820HQ CPU @ 2.70GHz, skipped
Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable
-------------------------------------------------------------
* Hash-Mode 22000 (WPA-PBKDF2-PMKID+EAPOL) [Iterations: 4095]
-------------------------------------------------------------
Speed.#01........: 63899 H/s (79.85ms) @ Accel:64 Loops:256 Thr:256 Vec:1
Speed.#02........: 7871 H/s (94.15ms) @ Accel:24 Loops:256 Thr:64 Vec:1
Speed.#*.........: 71771 H/s
Started: Wed Feb 11 03:43:33 2026
Stopped: Wed Feb 11 03:43:45 2026
Hashcat on Ubuntu 24.04.3 LTS benchmarked 71771 hashes per second.
Benchmarking Hashcat on Microsoft Windows 10 Pro π
On Microsoft Windows 10 pro, I run hashcat.exe -I to see all supported devices:
PS C:\Users\abanoub\Downloads\hashcat-7.1.2> .\hashcat.exe -I
hashcat (v7.1.2) starting in backend information mode
nvmlDeviceGetFanSpeed(): Not Supported
CUDA Info:
==========
CUDA.Version.: 12.8
Backend Device ID #01 (Alias: #02)
Name...........: Quadro M1200
Processor(s)...: 5
Preferred.Thrd.: 32
Clock..........: 1148
Memory.Total...: 4095 MB
Memory.Free....: 3411 MB
Memory.Unified.: 0
Local.Memory...: 48 KB
PCI.Addr.BDFe..: 0000:01:00.0
OpenCL Info:
============
OpenCL Platform ID #1
Vendor..: NVIDIA Corporation
Name....: NVIDIA CUDA
Version.: OpenCL 3.0 CUDA 12.8.98
Backend Device ID #02 (Alias: #01)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: Quadro M1200
Version........: OpenCL 3.0 CUDA
Processor(s)...: 5
Preferred.Thrd.: 32
Clock..........: 1148
Memory.Total...: 4095 MB (limited to 1023 MB allocatable in one block)
Memory.Free....: 3411 MB
Memory.Unified.: 0
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 573.71
PCI.Addr.BDF...: 01:00.0
OpenCL Platform ID #2
Vendor..: Intel(R) Corporation
Name....: Intel(R) OpenCL HD Graphics
Version.: OpenCL 3.0
Backend Device ID #03
Type...........: GPU
Vendor.ID......: 8
Vendor.........: Intel(R) Corporation
Name...........: Intel(R) HD Graphics 530
Version........: OpenCL 3.0 NEO
Processor(s)...: 8
Preferred.Thrd.: 7
Clock..........: 1050
Memory.Total...: 6505 MB (limited to 1626 MB allocatable in one block)
Memory.Free....: 3252 MB
Memory.Unified.: 1
Local.Memory...: 64 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 31.0.101.2111
Then I benchmarked how many hashes per second with this command .\hashcat.exe -b -m 22000:
PS C:\Users\abanoub\Downloads\hashcat-7.1.2> .\hashcat.exe -b -m 22000
hashcat (v7.1.2) starting in benchmark mode
Benchmarking uses hand-optimized kernel code by default.
You can use it in your cracking session by setting the -O option.
Note: Using optimized kernel code limits the maximum supported password length.
To disable the optimized kernel code in benchmark mode, use the -w option.
Kernel ./OpenCL/m22000-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel
nvmlDeviceGetFanSpeed(): Not Supported
CUDA API (CUDA 12.8)
====================
* Device #01: Quadro M1200, 3411/4095 MB, 5MCU
OpenCL API (OpenCL 3.0 CUDA 12.8.98) - Platform #1 [NVIDIA Corporation]
=======================================================================
* Device #02: Quadro M1200, skipped
OpenCL API (OpenCL 3.0 ) - Platform #2 [Intel(R) Corporation]
=============================================================
* Device #03: Intel(R) HD Graphics 530, 3252/6505 MB (1626 MB allocatable), 8MCU
Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable
-------------------------------------------------------------
* Hash-Mode 22000 (WPA-PBKDF2-PMKID+EAPOL) [Iterations: 4095]
-------------------------------------------------------------
Speed.#01........: 64867 H/s (87.99ms) @ Accel:9 Loops:512 Thr:1024 Vec:1
Speed.#03........: 7590 H/s (94.88ms) @ Accel:24 Loops:256 Thr:64 Vec:1
Speed.#*.........: 72456 H/s
Started: Wed Feb 11 02:40:05 2026
Stopped: Wed Feb 11 02:40:17 2026
Hashcat on Microsoft Windows 10 pro benchmarked 72456 hashes per second.
Crunching Hashcat numbers π
Hashcat benchmarked higher on Microsoft Windows 10 pro with a score of 72456 H/s. But on Ubuntu 24.04 LTS, Hashcat benchmarked 71771 H/s.
Microsoft Windows 10 outperformed Ubuntu 24.04 LTS by ~1% ((72456-71771)/71771 = 0.0095442449) despite having older CUDA version.
Takeaways π
If your laptop computer has an NVIDIA graphics card, it is safer to go with Microsoft Windows for the better optimizations of proprietary Nvidia driver. Or you may wait on the new opensource driver by Nvidia which is announced last year (2025).
If your laptop does not have an NVIDIA GPU, go with the latest Ubuntu LTS version; Ubuntu 26.04 LTS ( or Ubuntu 24.04 LTS if 26.04 is not yet available).
If you have Intel or AMD graphics card, you’ll get nearly exact performance on both Ubuntu Linux and Microsoft Windows. And sometimes Ubuntu Linux performs better with this hardware.
I hope you enjoyed reading this post as much as I enjoyed writing it. If you know a person who can benefit from this information, send them a link of this post. If you want to get notified about new posts, follow me on YouTube , Twitter (x) , LinkedIn , and GitHub .