|
the versions he's testing are like 3 updates apart so it's entirely possible that the regression isn't even due to 11.2.2, if it exists at all.
|
# ? Jan 11, 2018 02:09 |
|
|
# ? Jun 8, 2024 15:07 |
|
I think more people would be screaming if it’s that bad wouldn’t geekbench’s database be the best place to check?
|
# ? Jan 11, 2018 03:00 |
|
movax posted:I think more people would be screaming if it’s that bad a cursory glance of recent results bounces between slightly slower and drastically, so i wonder about what background processing might be happening after an update https://browser.geekbench.com/v4/cpu/6311128 https://browser.geekbench.com/v4/cpu/6311121 https://browser.geekbench.com/v4/cpu/6311084 https://browser.geekbench.com/v4/cpu/6311039 and then theres ones from just before that are terrible so: https://browser.geekbench.com/v4/cpu/6007671
|
# ? Jan 11, 2018 03:56 |
|
Updates often trigger a re-index which will jam up a core for a while (almost like that's 50% of your phones processing power 🤔)
|
# ? Jan 11, 2018 15:15 |
|
Stereotype posted:A lot of those look pretty "daily use" to me. JPEG, HTML5 parsing, PDF rendering, all down like 40%. seems like any camera accelerator worth a poo poo would have hardware jpeg encoding, that shouldn't drop off at all due to cpu performance. at least every camera accelerator i have tested has had it, and i have worked on 3 of them.
|
# ? Jan 11, 2018 15:25 |
|
genius yospos solution to avoiding the spectre bug is to get a phone with an operating system that will never get updated against the spectre bug
|
# ? Jan 11, 2018 16:24 |
|
|
# ? Jan 11, 2018 16:31 |
|
cross-postin:James Baud posted:In Intel news, there's also this:
|
# ? Jan 12, 2018 19:38 |
|
also, intel AMT is so much garbage: https://twitter.com/FSecure/status/951738561535119360 just turn that poo poo off, i bet this isn't even close to the last security hole in it
|
# ? Jan 12, 2018 19:39 |
|
did the iOS update, have noticed literally no difference
|
# ? Jan 13, 2018 03:01 |
|
yes, 11.2.2 has the patch haven’t noticed any perf issues even on my lovely old ipad
|
# ? Jan 13, 2018 03:03 |
|
a 50% loss still makes it way beyond any anroid(lol)
|
# ? Jan 14, 2018 13:56 |
|
BangersInMyKnickers posted:The sample size is 1 so maybe reserve judgement for a little bit i tried geekbench 3 on my iphone 7 and it's identical to the original performance of 10.0. i think something else is going on with that idiots phone, or maybe it only affects older iphones. it probably thermal throttled after sitting there updating to the new ios for 20 minutes and then he immediately ran the benchmark, or it was doing some background post-update housekeeping tasks while it was running. whatever it is, that article is fud. Hugh G. Rectum fucked around with this message at 09:16 on Jan 16, 2018 |
# ? Jan 16, 2018 09:13 |
|
https://twitter.com/johnregehr/status/953362475159732224
|
# ? Jan 16, 2018 22:26 |
|
Millcpu.txt
|
# ? Jan 17, 2018 09:17 |
|
BULLSHIT. Have you _looked_ at the patches you are talking about? You should have - several of them bear your name. The patches do things like add the garbage MSR writes to the kernel entry/exit points. That's insane. That says "we're trying to protect the kernel". We already have retpoline there, with less overhead. So somebody isn't telling the truth here. Somebody is pushing complete garbage for unclear reasons. Sorry for having to point that out. If this was about flushing the BTB at actual context switches between different users, I'd believe you. But that's not at all what the patches do. As it is, the patches are COMPLETE AND UTTER GARBAGE. They do literally insane things. They do things that do not make sense. That makes all your arguments questionable and suspicious. The patches do things that are not sane. WHAT THE F*CK IS GOING ON? And that's actually ignoring the much _worse_ issue, namely that the whole hardware interface is literally mis-designed by morons.
|
# ? Jan 22, 2018 04:58 |
|
Breakfast All Day posted:BULLSHIT. source your quotes
|
# ? Jan 22, 2018 07:32 |
|
NEED MORE MILK posted:source your quotes https://lkml.org/lkml/2018/1/21/192
|
# ? Jan 22, 2018 07:35 |
|
i wonder what linus’s conversations with intel have been like
|
# ? Jan 22, 2018 07:50 |
|
hobbesmaster posted:i wonder what linus’s conversations with intel have been like Probably very heated. Oh to be a fly on the wall in Linus Torvald's office when the Meltdown and Spectre vulnerabilities were revealed.
|
# ? Jan 22, 2018 08:14 |
|
linus posted:That's part of the big problem here. The speculation control cpuid
|
# ? Jan 22, 2018 11:15 |
|
Too much effort, they're deprecated CPUs anyways, and we can now hold a gun to people's heads about security and jack the price of our current gen which is compatible with our patches, and next gen procs through the roof!
|
# ? Jan 22, 2018 11:54 |
|
*looks at amd*
|
# ? Jan 22, 2018 12:00 |
|
orange juche posted:Too much effort, they're deprecated CPUs anyways, and we can now hold a gun to people's heads about security and jack the price of our current gen which is compatible with our patches, and next gen procs through the roof! atomicthumbs posted:*looks at amd* two potentially valid business strategies right there
|
# ? Jan 22, 2018 13:07 |
|
orange juche posted:Probably very heated. Oh to be a fly on the wall in Linus Torvald's office when the Meltdown and Spectre vulnerabilities were revealed. hitler_downfall_bunker_rant.wmv
|
# ? Jan 22, 2018 16:25 |
|
Malloc Voidstar posted:lol if intel refuses to ever patch spectre spectre doesn’t seem like a very practical attack at the moment at least
|
# ? Jan 22, 2018 17:26 |
|
Best Bi Geek Squid posted:hitler_downfall_bunker_rant.wmv tbf from the quotes i've seen in the past that strikes me as a pretty regular phenomenon with ol' linus
|
# ? Jan 22, 2018 17:42 |
|
actually, wait. it's not quite downfall_rant.wmv because linus wouldn't ask people to leave the room
|
# ? Jan 22, 2018 17:44 |
|
intel is going to make the spectre fix xeon only like every other basic and sane feature over the past decade
|
# ? Jan 22, 2018 18:00 |
|
Farmer Crack-rear end posted:actually, wait. it's not quite downfall_rant.wmv hey nvidia, you got outclassed in being egregious shitheads today! go get yourself an ice cream cone!
|
# ? Jan 22, 2018 18:00 |
|
the only reasonable fix for these defects is to recall every affected processor and provide replacements for no cost
|
# ? Jan 22, 2018 19:24 |
|
Stymie posted:the only reasonable fix for these defects is to recall every affected processor and provide replacements for no cost for meltdown, maybe. meltdown was kind of an obvious fuckup that intel really, really, really should not have made. that one is on them. spectre is a much more subtle problem that affects everyone industry-wide
|
# ? Jan 22, 2018 19:26 |
|
death to the Thinking Machines
|
# ? Jan 22, 2018 20:23 |
|
dude Thinking Machines died in 1994 and was bought by Sun
|
# ? Jan 22, 2018 20:30 |
|
oh lol I didn't know that was a real thing. I was talking dune
|
# ? Jan 22, 2018 20:56 |
|
My Linode VPSes got patched for Meltdown or whatever and performance seems identical but for some reason it's now reporting way more I/O
|
# ? Jan 22, 2018 21:55 |
|
my motherboard vendor only has a beta bios patch up i try to avoid things with "beta bios" in them
|
# ? Jan 22, 2018 22:12 |
|
HAIL eSATA-n posted:my motherboard vendor only has a beta bios patch up how to brick your machine in four clicks or less!
|
# ? Jan 22, 2018 22:31 |
|
Linguica posted:My Linode VPSes got patched for Meltdown or whatever and performance seems identical but for some reason it's now reporting way more I/O Same here, I love the weird artifacts on IPv6 monitoring:
|
# ? Jan 23, 2018 11:16 |
|
|
# ? Jun 8, 2024 15:07 |
|
Linguica posted:My Linode VPSes got patched for Meltdown or whatever and performance seems identical but for some reason it's now reporting way more I/O hrm, wonder if this is just an artifact of the trampoline mapping in of the kernel not being excluded from io (likely happening on a slightly different path than usual) despite already existing in (unmapped) physical memory? can't think why there would be additional actual blocks io'ed e: otoh my very lightly, but consistently, loaded virtual machine has other io weirdness which is partially consistent with yours, so there may be something else going on: the start of the calm period also coincides with a reboot, so i wonder if it isn't just some linode/kernel change Cybernetic Vermin fucked around with this message at 12:23 on Jan 23, 2018 |
# ? Jan 23, 2018 12:16 |