F5F Stay Refreshed Hardware Desktop Evaluate machine processing capability by assessing its computational speed and performance metrics.

Evaluate machine processing capability by assessing its computational speed and performance metrics.

Evaluate machine processing capability by assessing its computational speed and performance metrics.

T
TheTerrawr
Junior Member
3
12-13-2023, 08:24 PM
#1
Hey! I’m trying to gauge the CPU performance of various laptops and servers for your analysis. To start, it seems challenging to pinpoint the optimal setup without running tests on each device with all your algorithms. You mentioned using sorting, ML training, and some complex research tasks—some single-threaded, others multi-threaded. Since you’re working across macOS, Windows, and UNIX systems, the results should reflect a broad range of environments. I’d suggest splitting your tests into single-threaded and multi-threaded scenarios to compare performance directly. Tools like LinPack could help with precise benchmarking, especially for math-heavy workloads. If you need deeper insights, consider logging detailed metrics during execution rather than relying solely on benchmarks. Let me know if you’d like guidance on structuring those tests!
T
TheTerrawr
12-13-2023, 08:24 PM #1

Hey! I’m trying to gauge the CPU performance of various laptops and servers for your analysis. To start, it seems challenging to pinpoint the optimal setup without running tests on each device with all your algorithms. You mentioned using sorting, ML training, and some complex research tasks—some single-threaded, others multi-threaded. Since you’re working across macOS, Windows, and UNIX systems, the results should reflect a broad range of environments. I’d suggest splitting your tests into single-threaded and multi-threaded scenarios to compare performance directly. Tools like LinPack could help with precise benchmarking, especially for math-heavy workloads. If you need deeper insights, consider logging detailed metrics during execution rather than relying solely on benchmarks. Let me know if you’d like guidance on structuring those tests!

T
TrueMCcrafter
Member
67
12-13-2023, 09:41 PM
#2
Would you like to track the overall system power use or focus only on the CPU's energy draw? For the first step, consider using a watt meter. If you prefer not to handle hardware modifications, software monitoring is a viable alternative, though it may not be as precise.
T
TrueMCcrafter
12-13-2023, 09:41 PM #2

Would you like to track the overall system power use or focus only on the CPU's energy draw? For the first step, consider using a watt meter. If you prefer not to handle hardware modifications, software monitoring is a viable alternative, though it may not be as precise.

Y
Yteews76
Junior Member
10
12-14-2023, 09:49 AM
#3
I think OP is talking about stress testing and performance rather than actual power. Here are some suggestions that work across Unix, macOS, and Windows environments.
Y
Yteews76
12-14-2023, 09:49 AM #3

I think OP is talking about stress testing and performance rather than actual power. Here are some suggestions that work across Unix, macOS, and Windows environments.

1
193over71
Member
169
12-15-2023, 12:10 AM
#4
Consider using cross-platform tools that mimic real-world tasks. I’d start by testing 7-Zip, running benchmarks with multiple cores and then restricting it to a single thread.
1
193over71
12-15-2023, 12:10 AM #4

Consider using cross-platform tools that mimic real-world tasks. I’d start by testing 7-Zip, running benchmarks with multiple cores and then restricting it to a single thread.

I
ipad_100
Junior Member
29
12-18-2023, 01:15 PM
#5
Sure, the count of computations wasn't included since I wasn't certain it would fully address all my requirements.
I
ipad_100
12-18-2023, 01:15 PM #5

Sure, the count of computations wasn't included since I wasn't certain it would fully address all my requirements.

L
LooseDawg
Senior Member
628
12-20-2023, 12:52 AM
#6
Great idea!
L
LooseDawg
12-20-2023, 12:52 AM #6

Great idea!