|
Measly Twerp posted:I've heard this a lot, but to be honest I've not seen it.
|
# ? Nov 25, 2017 17:55 |
|
|
# ? May 15, 2024 03:59 |
|
systemd is a scourge and needs to be purged from the Linux ecosystem with maximum prejudice.
|
# ? Nov 25, 2017 18:15 |
|
Paul MaudDib posted:Why not mount kernel internals r/w to userspace? Surely accidents are never going to happen, anyone who wants to write to that will know what they're doing. Right? They do. Is under the /proc filesystem. You can read them, you can write to them and is the kernel module's job to ensure the system doesn't panic or gets bricked. Look, nobody likes Leonard and that's fine. But there are enough systemd problems to point at without having to bring up this.
|
# ? Nov 25, 2017 18:41 |
|
Obsurveyor posted:systemd is a scourge and needs to be purged from the Linux ecosystem with maximum prejudice. I have to say I am coming around to this view. It feels over engineered and hides too much stuff from what I've seen from it.
|
# ? Nov 25, 2017 21:16 |
|
NewFatMike posted:Sorry, nerds, there's a little hardware news today: Bringing this over to the new page since it got sniped
|
# ? Nov 25, 2017 22:23 |
|
Arzachel posted:It's just market segmentation, there are no manufacturing defects that would break SMT specifically. Do they disable it via some kind of code or physical change?
|
# ? Nov 25, 2017 22:40 |
|
Darth Llama posted:Do they disable it via some kind of code or physical change? There are polysilicon fuses that control core configuration and whether various features are enabled. So to make an R3 you take the basic 8C16T processor and blow the fuses that turn off 4 of the cores and the hyperthreading. Sometimes there are mistakes or omissions in the configuration, for example this is how you get 8C8T R3s and such. Intel's E5-1600 processor line also happened to use virtually the same feature fuse configuration as the consumer i7s and so they actually had the multiplier unlocked and were overclockable up to the v3 generation.
|
# ? Nov 25, 2017 22:46 |
|
Paul MaudDib posted:There are polysilicon fuses that control core configuration and whether various features are enabled. So to make an R3 you take the basic 8C16T processor and blow the fuses that turn off 4 of the cores and the hyperthreading. Sometimes there are mistakes or omissions in the configuration, for example this is how you get 8C8T R3s and such. Intel's E5-1600 processor line also happened to use virtually the same feature fuse configuration as the consumer i7s and so they actually had the multiplier unlocked and were overclockable up to the v3 generation. I'd be laughing so hard if in order to guarantee that the fuses don't accidentally gently caress your chip up, they overenginner them so hard that occasionally they don't respond to the 'blow me up please' command, which is why we're seeing occasional R3s with 8 cores or whatever.
|
# ? Nov 26, 2017 14:23 |
id be surprised if it wasn't something on the order of 'fuse exists in the main power conduit to the core' that they just run a high current through using a couple of special pads that don't get connected to anything else, so if something did manage to blow the fuse in a customer computer, i'd be something that'd hosed it up in other ways as well. Of course, that'd also fall victim to mfg'ing errors too, so conceivably a bad mask or whatever could let through a whole batch of chips which couldn't shut off.
|
|
# ? Nov 26, 2017 17:27 |
|
More developments post Ryzen NPT tables fix: Headless. Guest VFIO. https://www.youtube.com/watch?v=1MI1s4hZ_yE
|
# ? Nov 27, 2017 06:21 |
|
This is freaking awesome. I hope NVidia gets it's head out of it's rear end regarding VMs.
|
# ? Nov 27, 2017 10:29 |
|
I got my replacement Ryzen from AMD today, week 39 processor. I can confirm they didn't test it since both seals are unbroken. Also, they just sent another full retail CPU with the Wraith Spire, even though they told me to just ship back the CPU without the cooler. So now I have an extra cooler I guess.
|
# ? Nov 27, 2017 21:00 |
|
SwissArmyDruid posted:More developments post Ryzen NPT tables fix: also the patch for the long standing "ignore cache setting on AMD"(i.e. NPT bug) was accepted into 4.15
|
# ? Nov 27, 2017 21:45 |
|
APU Benches and review! This one is from The Tech Report on the R5 2500U https://techreport.com/review/32877/amd-ryzen-5-2500u-apu-reviewed TL;DR: Still has a weird memory latency thing, like almost twice the delay as some Intel processors. I have no idea why that is since everything is on a single CCX. Largely neck-and-neck with the i7 8250U - in most tests the Intel chip is a little bit ahead, some a lot a bit ahead. Looks like the largely Broadwell level of performance is holding up. Apparently it's bad for audio processing benchmarks? I'm no expert, but it definitely fared worse than Intel's offerings. My favorite part, the gaming: 720p 30+FPS seems to be the name of the game. Between 30-40FPS is what we're getting in titles like The Witcher 3, which is really, really cool. Getting that level of performance out of a 15W thermal envelope has me really excited for some of the beefier SKUs. The article compared it to a system equipped with the MX 150 the whole time, but that GPU alone has higher power draw than this APU. I'm super excited for the R5 2700U to start showing up and to see what the R7 units will look like. Hopefully we'll hit 900p 30+ FPS in one of those mobile SKUs. It's really nice to see an impressive, let alone competent, product out of AMD's APU division.
|
# ? Nov 28, 2017 04:39 |
|
The fact that it's touching 1080p 60fps at low in some things is actually what has me excited. 2700U is going to be cool. Not quite MX150 territory but it's also at a much lower TDP. Weird about the latency thing. Any thoughts? Audio processing in particular is known to be really latency-sensitive so that one's not surprising given the prior.
|
# ? Nov 28, 2017 05:05 |
|
Paul MaudDib posted:The fact that it's touching 1080p 60fps at low in some things is actually what has me excited. 2700U is going to be cool. Not quite MX150 territory but it's also at a much lower TDP. Willing to bet that it's simply a BIOS issue.
|
# ? Nov 28, 2017 08:46 |
|
Paul MaudDib posted:but a random one-liner should not be able to brick your system, full stop. Poettering is a for even suggesting it, even by Linus standards. especially by Linus standards, even. is that why amdflash.exe can brick any system any time? it's just a random one liner that needs to be run as root
|
# ? Nov 28, 2017 09:51 |
|
NewFatMike posted:APU Benches and review! This one is from The Tech Report on the R5 2500U I'm glad to see that the HP Envy decided that yes, dual-channel was actually worth it, even if it is only 2x4GB.
|
# ? Nov 28, 2017 18:45 |
|
Paul MaudDib posted:Why not mount kernel internals r/w to userspace? Surely accidents are never going to happen, anyone who wants to write to that will know what they're doing. Right? ...you mean like /dev/mem?
|
# ? Nov 28, 2017 21:51 |
|
Truga posted:is that why amdflash.exe can brick any system any time? it's just a random one liner that needs to be run as root and running rm -rf / not only requires sudo do to any major damage, but unless you have some ancient version you need to explicitly add --no-preserve-root to the command line you are probably more likely to append the wrong device to a dd command than to go through all the steps needed to recursively erase your entire hard drive with rm
|
# ? Nov 28, 2017 22:10 |
|
the real horror here is that some firmwares do not have any kind of "use defaults if config is missing", not that it's possible to delete the config if running rm -rf /
|
# ? Nov 28, 2017 23:04 |
|
SwissArmyDruid posted:I'm glad to see that the HP Envy decided that yes, dual-channel was actually worth it, even if it is only 2x4GB. Right? I think Lenovo are still slow on the uptake on that one.
|
# ? Nov 28, 2017 23:06 |
|
kujeger posted:the real horror here is that some firmwares do not have any kind of "use defaults if config is missing", not that it's possible to delete the config if running rm -rf /
|
# ? Nov 28, 2017 23:23 |
|
kujeger posted:the real horror here is that some firmwares do not have any kind of "use defaults if config is missing", not that it's possible to delete the config if running rm -rf / Yuuuuuuup. Leaving the defaults in NVRAM instead of including one with the firmware image is at least double hubris (both that your NVRAM is inviolable and that your shipment defaults will be valid for every firmware image you ever deploy).
|
# ? Nov 29, 2017 00:04 |
|
AMD keep posting special offers on AMD products from ocuk. Someone really needs to tell them ocuk are pro white nationalists.
|
# ? Nov 29, 2017 04:29 |
|
Raven Ridge will be coming to existing AM4 motherboards with BIOS update
|
# ? Nov 29, 2017 05:09 |
|
So sad news, despite having it as an option in the BIOS, having the IOMMU setting enabled causes Server 2016 to poo poo itself and refuse to load on basially every x399 motherboard. Which means no discrete device passthrough, and no shiny new tape drive directly attached to the file server VM Hopefully they patch it in another few months, otherwise I'm gonna be kinda sad.
|
# ? Nov 29, 2017 12:24 |
|
As in you have to flash the board with a different CPU first? feedmegin posted:...you mean like /dev/mem? Is that user-writeable by default?
|
# ? Nov 29, 2017 19:49 |
|
Munkeymon posted:Is that user-writeable by default? If you're root? Yes.
|
# ? Nov 30, 2017 13:58 |
|
feedmegin posted:If you're root? Yes. Right, I thought Paul's original example was making the BIOS user-writeable by non-root users, but I may have misunderstood.
|
# ? Nov 30, 2017 16:38 |
|
You need to be root to write to the efivars. edit: which is why this is so ridiculous. The EFI explicitly presents these vars as writable. So they get set up as writable by root. But (some) EFIs can't handle them being cleaned out. ...and for some insane reason people are personally blaming Poettering for this. kujeger fucked around with this message at 18:21 on Nov 30, 2017 |
# ? Nov 30, 2017 18:19 |
|
kujeger posted:You need to be root to write to the efivars. If I write to the boot drive as root, the operating system needs to be reinstalled If I write to the EFIs as root, the computer has been turned into a brick There's a difference
|
# ? Nov 30, 2017 18:30 |
|
Should the os be modified to workaround bad hardware? Its called efivars no combination of variables or lack there of should make a system unbootable, the real solution is a firmware update.
|
# ? Nov 30, 2017 20:42 |
|
It's a stack of assholes saying "I just need to engineer this to work, that issue is someone else's problem" the whole way down
|
# ? Nov 30, 2017 20:51 |
|
Rastor posted:It's a stack of assholes saying "I just need to engineer this to work, that issue is someone else's problem" the whole way down TBCF this is why Pottering is "blamed" for this since that's basically his answer too.
|
# ? Nov 30, 2017 22:21 |
|
Rastor posted:It's a stack of assholes saying "I just need to engineer this to work, that issue is someone else's problem" the whole way down The first line of defense against bricking your system is the firmware itself. If that fails, then the responsibility falls on the kernel (which is where it was fixed anyway, since it is much harder to make firmware vendors hire anything else but drunken monkeys). The responsibility never falls on the userland, doesn't matter if it's pid 1 or 10000.
|
# ? Nov 30, 2017 22:30 |
|
Turns out all those 2500U benchmarks actually were using a 25W TDP due to mobile XFR. Guess that explains the difference in sustained performance vs the 8250U that people were noticing.
|
# ? Dec 1, 2017 19:44 |
|
Paul MaudDib posted:Turns out all those 2500U benchmarks actually were using a 25W TDP due to mobile XFR. I was just about to post the same thing. Seems to explain the battery life stuff going on as well. At least HP seem to be giving it its best in their implementation (HDD as the default option notwithstanding) - dual channel RAM, non garbage screen, and mobile XFR for a little extra go juice. Mixed blessing, I guess, but with all the fairly rapid improvements on desktop Ryzen with power management and things like that, there's probably a few easy pickings for upcoming updates. I was kinda schadenfreude hoping that resource killing telemetry HP secretly installed in everyone's laptop was going to bork reviews for everyone, but I guess the comedy timeline we live in only goes so far.
|
# ? Dec 1, 2017 21:41 |
|
Graphics processing costs watts, yo. Always has, always will.
|
# ? Dec 2, 2017 05:24 |
|
|
# ? May 15, 2024 03:59 |
|
That is kind of disappointing honestly, I was hoping for a hard 15w clamp but given how the Envy is built (active cooling and all) it's probably what made sense to go with. I am really interested in seeing a true 15w and 9w test.
|
# ? Dec 2, 2017 06:22 |