Windows 10 Outperforming Linux On A ~$5000 Laptop, Ubuntu Beating Clear Linux

[Michael]

Michael is "-mbranches-within-32B-boundaries" used by all distros or only on CL? it is set as default buildflag on CL - it concerns the JCC mitigation https://www.intel.com/content/dam/su...de-erratum.pdf
But this is just speculation.

That helps the performance in light of JCC Erratum on supported assemblers, but note that Clear still came in last with these tests, so doesn't explain the difference overall.

https://www.phoronix.com/scan.php?page=article&item=intel-jcc-microcode&num=1

[blacknova]

How loud is this laptop?
I'm looking into getting new one DTR and I definitely want to get a silent one. My current DTR - HP8770w can be really loud under load.
If I'll be able to find DTR with Ryzen 4k APU I'll be especially happy

[Michael]

How loud is this laptop?
I'm looking into getting new one DTR and I definitely want to get a silent one. My current DTR - HP8770w can be really loud under load.
If I'll be able to find DTR with Ryzen 4k APU I'll be especially happy

Pretty quiet - basically silent and idle and just hearing the fan ramp up under multi-threaded load, I don't have any complaints over the noise.

[CochainComplex]

That helps the performance in light of JCC Erratum on supported assemblers, but note that Clear still came in last with these tests, so doesn't explain the difference overall.

https://www.phoronix.com/scan.php?pa...icrocode&num=1

it only helps if the Microcode supports it otherwise it has an adverse effect....JCC Documents says Coffe Lake based Xeon E should be covered. So yes but if there is something wrong opposite outcome... Usage of more avx/avx2 instructions? downclocking the cpu ending up with slower total speed?

Im just thinking through different possibilities ...because this simply does not fit to the normal Clear Linux numbers on recent intel CPU's.
I have also a Xeon 2286M (dell precision 7490) and I'm running on Clear Linux - am I wasting peformance ?

[AdamOne]

Really useful test results, Phoronix gets thumbs UP

[Pyth0n]

My first thought on HP laptop misbehaving under non-Windows: they screwed-up ACPI. Again. Convince me not on-purpose.

[Michael]

it only helps if the Microcode supports it otherwise it has an adverse effect....JCC Documents says Coffe Lake based Xeon E should be covered. So yes but if there is something wrong opposite outcome... Usage of more avx/avx2 instructions? downclocking the cpu ending up with slower total speed?

Im just thinking through different possibilities ...because this simply does not fit to the normal Clear Linux numbers on recent intel CPU's.
I have also a Xeon 2286M (dell precision 7490) and I'm running on Clear Linux - am I wasting peformance ?

Clear does ship with the updated microcode.

[CochainComplex]

Clear does ship with the updated microcode.

I know - I meant more issues on GNU Assembler + (inside) Microcode it self side.

[creative]

While I don't use laptops and only have one desktop, its still nice to see Manjaro performing strong. I knew it was faster than in past benchmarks. Been a blazing fast distribution for me.

[Volta]

Without data about mitigations on Windows it doesn't tell much. Clear Linux losing it's another strange thing. CPU governor scaling problems?