I wasted 3 whole hours benchmarking my intel CPU, with mitigations enabled/disabled. There was no performance difference on any cpu/ram benchmark. Why the fuck are there articles talking about this supposed 40% performance hit? I was expecting to get 40% performance boost by disabling the shit. But nothing. Zero. 0% performance gained. Why waste people's time like this? Fuck you shills
processor: i7 4820k from 2013.
Never falling for Jow Forums memes again
Evan Phillips
Other urls found in this thread:
venturebeat.com
twitter.com
Ian Cook
ITT: user don't know how to turn on/off migrations.
Julian Scott
>Why the fuck are there articles talking about this supposed 40% performance hit?
they're up to 40% under some workloads that generally don't affect normal desktop usage or gamer faggotry, cloud/vps companies in general are absolutely seething because of the constant intel specific vulnerabilities
also
Jordan King
meh it's more ~15-20%ish
Josiah Peterson
RETARD
Matthew Fisher
/thread
Jason Lewis
k
Aiden Kelly
I wonder who could be behind this psot
Jonathan Powell
The 40% was with hyperthreading disabled.....
Go ahead and turn that off, now youll see a big difference.
Blake Baker
/thread
Gabriel Torres
Are you repeating the same test that was done in the article you question?
Are you you you're test actually tests the features with expected performance decrease?
Nicholas Foster
>migrations
kek
Nathan Ortiz
the articles say nothing about the benchmarking software or use case they tested it on and it's infuriating.
>venturebeat.com
then I did more research and turns out this '40%' is due to retards turning off multithreading. why would any sane person do that?
Michael Ward
forgot to quote
read & pls respond
Adrian Reyes
that's nasty. those jews made the article implying that the PATCH itself caused a 40% drop. but the 40% drop is from braindeads disabling hyperthreading
Landon Cooper
>why would any sane person do that?
Hyperthreading is the vulnerability.
Asher Perry
No one is turning off hyperthreading.
You can if you want to stop that one specific thing but its hardly worth it for anyone but NSA tier government hardware.
Parker Reed
>posts desktop benchmark tools to disprove performance loss for virtualisation providers
Leo Harris
Go benchmark the SSD, faggot, you'll see
James Morris
I use virtualization on my servers at home (no, not for shit Jow Forumseddit memelab tier shit that gets posted here, I actually goldfarm gook MMOs and make $). Is there a performance hit if I keep hyperthreading enabled?
I actually benchmarked my RAM before/after mitigations, same score.
Will try SSD and report back
Ethan Edwards
>Once these updates are applied, it may be appropriate for some customers to consider additional steps. This includes customers who cannot guarantee that trusted software is running on their system(s) and are using Simultaneous Multi-Threading (SMT). In these cases, customers should consider how they utilize SMT for their particular workload(s), guidance from their OS and VMM software providers, and the security threat model for their particular environment.
Nolan Lewis
the vulnerability (and many of the other similar spectre type vulns) are that malicious or untrusted programs may be able to read the memory of other processes on the same core/of the same process(think malicious sites reading memory from your browser)/etc, if you're only running software you trust and would run on your regular computer but are virtualising it to run more of it and aren't hosting programs for other people you'll be fine without disabling hyperthreading
Tyler Flores
Also, what is the fucking point of having a 4820K instead of a 4770K
Asher Gomez
you do know that you need to hard reboot between switches and tests, right? RIGHT, mr shnozenstein?
Lincoln Wood
and to answer your question specifically it depends entirely on your workload, many of the worst spectre mitigations with 10% perf loss only affected programs which relied on memory latency and things like that and most users/servers weren't affected
Ayden Ward
>Jeremy Horowitz
Into the trash it goes.
Gavin Sanchez
negligible difference
yes, been doing it every time.
its my first pc i ever built so I wanted to go with what was the most high end at the time and some benchmarks for this cpu looked better than 4770k's for minimal extra monies.
thanks for the info frens. I guess I fell for the Jow Forums hype from false flaggers and the clickbait on youtube about this with basically no real world proof on where a real person would be affected by this
yeah... I should have noticed (((that))).
Lincoln Parker
At the time it might have been silly, but these days having an X79 platform is a fair bit nicer than being stuck with a Z77.
Joseph Cook
You didn't do it right, faggot.
Michael Russell
This. Actually it’s a sloppy Intel shill.
David Adams
>braindeads
We know there are many more vulns in Intel’s SMT implementations just waiting to be discovered.
Jaxon Thomas
>SATA SSD
No wonder. Try the NVMe mod and run your tests again.
Nolan Cooper
lmao you clearly didn't do anything
Mason Williams
Wait so this is just for nvme ssds? i dont understand how this warrants such outrageous claims like 20% cpu performance drop when nobody uses nvme other than people who built recently (either highschoolers or people coming off 2008 builds) because lga2011 procs are still fine til 2025+