|
Sinestro posted:I decided to wait until today, since it would have needed massive retooling when the embargo dropped anyway. Should be tonight, I just need to put things together when I get home from work, I know what I want to say. SourKraut posted:What time is the embargo expected to drop? My bad about the embargo, people are saying it's actually on release day (Thursday March 2). I was hearing February 28 but that must have been confusion caused by the "Capsaicin and Cream" GPU event.
|
#
?
Feb 28, 2017 22:31
|
|
|
|
| # ? May 21, 2024 22:57 |
|
|
Ah, that's nice! So yeah, it'll be March 2nd, which is what I thought before.
|
|
#
?
Feb 28, 2017 23:11
|
|
|
Rastor posted:My bad about the embargo, people are saying it's actually on release day (Thursday March 2). I was hearing February 28 but that must have been confusion caused by the "Capsaicin and Cream" GPU event. Sinestro posted:Ah, that's nice! So yeah, it'll be March 2nd, which is what I thought before. Thanks! Though now I am suspicious, since it seems like if they were ready to really be excited about Ryzen, they'd let the NDA lift before the actual release date.
|
|
#
?
Mar 1, 2017 00:46
|
|
|
the errata seems to be significant enough apparently
|
|
#
?
Mar 1, 2017 01:08
|
|
|
Platystemon posted:Here’s an article where someone actually successfully tracks down a single bit error in some code. That strikes me as a computer autist that doesn't know bad RAM when he sees it. But your point remains.
|
|
#
?
Mar 1, 2017 01:45
|
|
|
5.2 GHz on LN2. https://www.pcper.com/news/Processors/Overclockers-Push-Ryzen-7-1800X-52-GHz-LN2-Break-Cinebench-Record
|
|
#
?
Mar 1, 2017 03:08
|
|
|
SwissArmyDruid posted:5.2 GHz on LN2. quote:In addition to the exotic LN2 cooling, the Ryzen 7 1800X needed 1.875 volts to hit 5.20 GHz. That 5.20 GHz was achieved by setting the base clock at 137.78 MHz and the multiplier at 37.75. Using these settings, the chip was even stable enough to benchmark with a score of 2,363 on Cinebench R15’s multi-threaded test. But how does it perform attached to a passive radiator in spaaaaaaace? WTFTECH: SpaceX announces 2018 OC tourism in free return orbit around dark side of moon, clocks promised to be "out of this world" Potato Salad fucked around with this message at 03:52 on Mar 1, 2017 |
|
#
?
Mar 1, 2017 03:47
|
|
|
Radiation is by far the slowest method of heat transfer; space would be awful.
|
|
#
?
Mar 1, 2017 04:51
|
|
|
Potato Salad posted:But how does it perform attached to a passive radiator in spaaaaaaace? poo poo: in space, you can't conduct heat away since there's no atmosphere with which to do so. Radiation is terribly inefficient by comparison.
|
|
#
?
Mar 1, 2017 04:51
|
|
|
turn left hillary!! noo posted:Radiation is by far the slowest method of heat transfer; space would be awful. Nam Taf posted:poo poo: in space, you can't conduct heat away since there's no atmosphere with which to do so. Radiation is terribly inefficient by comparison. While I'm not saying it'd work, how would it be different from the flat plate radiators present on spacecraft?
|
|
#
?
Mar 1, 2017 05:05
|
|
|
Well it says they reject heat at a rate of 100-350 W/m^2, and you can't fold it up like a normal heatsink, so it sounds bad... Edit: to be clear those would work, just poorly.
|
|
#
?
Mar 1, 2017 05:11
|
|
|
Would be pretty funny to have your space computer with two huge meter square wings of radiator material jutting out the back into space.
|
|
#
?
Mar 1, 2017 05:14
|
|
|
I think they have to watercool the 486 processors that run the International Space Station's core functions because space is such a poo poo place to cool things.
|
|
#
?
Mar 1, 2017 05:19
|
|
|
FuturePastNow posted:I think they have to watercool the 486 processors that run the International Space Station's core functions because space is such a poo poo place to cool things. Loss of cooling is one of the most urgent emergencies ISS can encounter.
|
|
#
?
Mar 1, 2017 05:33
|
|
|
FuturePastNow posted:I think they have to watercool the 486 processors that run the International Space Station's core functions because space is such a poo poo place to cool things. I know they use liquid ammonia to cool the solar panels, but looking around apparently each segment uses different coolants http://www.space.com/21059-space-station-cooling-system-explained-infographic.html  The laptops they use obviously are air cooled and that air is conditioned by the station, not sure about the core computer systems though.
|
|
#
?
Mar 1, 2017 05:44
|
|
|
FuturePastNow posted:I think they have to watercool the 486 processors that run the International Space Station's core functions because space is such a poo poo place to cool things. Wait, ye olde 486 processors? Suddenly Alien/s take on computer technology doesn't seem outlandish. It's all retro because it's cheap, sturdy and runs cold.
|
|
#
?
Mar 1, 2017 05:50
|
|
|
FaustianQ posted:Wait, ye olde 486 processors? Suddenly Alien/s take on computer technology doesn't seem outlandish. It's all retro because it's cheap, sturdy and runs cold. It's more like the stuff has to be certified rad-hardened
|
|
#
?
Mar 1, 2017 06:08
|
|
|
ISS uses 386. Hubble is on 486. Certification for flight takes a long rear end time. Also, reminder that the aerospace industry loves ADA. http://www.ada-auth.org/cpl/lists/CPLbase.html
|
|
#
?
Mar 1, 2017 06:09
|
|
|
Why not an ARM solution? Or is legacy code holding them back? God NASA needs an actual loving budget.
|
|
#
?
Mar 1, 2017 06:53
|
|
|
FaustianQ posted:Why not an ARM solution? Or is legacy code holding them back? God NASA needs an actual loving budget. FuturePastNow posted:It's more like the stuff has to be certified rad-hardened
|
|
#
?
Mar 1, 2017 07:03
|
|
|
Just use some old galaxy note 7's
|
|
#
?
Mar 1, 2017 07:07
|
|
|
Keep in mind we're not talking about computers used for scientific observation, but the guidance computers used for controlling the spacecraft. It's not exactly a computationally demanding application. Additionally, the guidance computers for ISS would have been designed in the 90s, and Hubble Space Telescupe was built in the 80s and intended for launch in 1986 before Challenger was destroyed.
Farmer Crack-Ass fucked around with this message at 07:14 on Mar 1, 2017 |
|
#
?
Mar 1, 2017 07:11
|
|
|
Well I assume time is a constraint, but I'm wondering if lack of funds aren't also slowing the process down.
|
|
#
?
Mar 1, 2017 07:19
|
|
FaustianQ posted:Why not an ARM solution? Or is legacy code holding them back? God NASA needs an actual loving budget. What they have works fine for the application it's being used in and is known to be stable which is much, much, much more important than how fast it is.
|
|
|
#
?
Mar 1, 2017 07:26
|
|
|
The 386 and 486 saw a lot of embedded use in the 90s so that's not even very atypical considering the time period.
|
|
#
?
Mar 1, 2017 07:32
|
|
|
MaxxBot posted:The 386 and 486 saw a lot of embedded use in the 90s so that's not even very atypical considering the time period. And 2000s. 386 production ended in 2007
|
|
#
?
Mar 1, 2017 07:38
|
|
|
There's no atmosphere to shield components from radiation, so they must have radiation resistant electronics. It would suck to have computational errors in guidance computers etc. when you are in space.
|
|
#
?
Mar 1, 2017 10:06
|
|
|
   1440p resolution but we have no idea what GPUs and graphic settingswere used for each of the slides. for all we know these could be pretty GPU bottlenecked. source edit: bonus XFR slide 
eames fucked around with this message at 10:37 on Mar 1, 2017 |
|
#
?
Mar 1, 2017 10:30
|
|
|
I know these are AMD's marketing benchmarks but I'm still trying to figure them out. The 1700 getting closer to the 7700K than the 1700X gets to the 6800K has me scratching my head and I can only conclude that it's a L3+Real Core advantage. Oh, wait, the 1700/7700K is using a 1070 while the 1700X/6800K is using a 1080? I guess that means the lower SKU comparison is much more GPU bound as well.
|
|
#
?
Mar 1, 2017 10:45
|
|
|
FuturePastNow posted:It's more like the stuff has to be certified rad-hardened It's this. This isn't even a time constraint on certifying things. A random gamma ray (and there are a lot in space) hitting the big 386 chip doesn't really have much chances of displacing something important. A high energy neutron hitting something in your modern <30nm arch is going to wreck it hard. 386 and 486 are perfect for that use, just due to the sheer size of each element.
|
|
#
?
Mar 1, 2017 10:52
|
|
|
FaustianQ posted:I know these are AMD's marketing benchmarks but I'm still trying to figure them out. The 1700 getting closer to the 7700K than the 1700X gets to the 6800K has me scratching my head and I can only conclude that it's a L3+Real Core advantage. It's bad by design, or cherry-picked if you will. Much like AMD, which is also bad by design.
|
|
#
?
Mar 1, 2017 10:53
|
|
|
Truga posted:It's this. This isn't even a time constraint on certifying things. A random gamma ray (and there are a lot in space) hitting the big 386 chip doesn't really have much chances of displacing something important. A high energy neutron hitting something in your modern <30nm arch is going to wreck it hard. 386 and 486 are perfect for that use, just due to the sheer size of each element. That works for now, but for the future, will we keep the 486 foundries running, stockpile all the rad‐hard chips we could every possibly need, or rad‐harden newer tech? Like, instead of having one processor where a gamma ray won’t do much, have a handful running in lock‐step with a bunch more as hot spares. When one of them disagrees with the consensus, restore it to a known good state. If it continues to err, remove it from the pool and bring a hot spare online.
|
|
#
?
Mar 1, 2017 11:37
|
|
|
There's already vastly more powerful cpus available for space use, but apart from some exceptions like curiosity, most deep space missions will still opt for the vastly cheaper ("only" $200k or so per) 3-486 era chips. Probably not least because due to the need for very simple code (you can't really debug critical bugs when your thing is several AU away) and very low bandwidths available at those distances, I imagine there's no need for faster CPUs on those.
|
|
#
?
Mar 1, 2017 12:15
|
|
|
SourKraut posted:While I'm not saying it'd work, how would it be different from the flat plate radiators present on spacecraft?
|
|
#
?
Mar 1, 2017 15:12
|
|
|
drat this thread is really going off the reservation in the final days before ryzen. You guys it's gonna be ok, you don't have to talk about CPUs on the moon just hang on a little longer!
|
|
#
?
Mar 1, 2017 15:15
|
|
|
I want to see Ryzen benchmarks compared to the ISS's 486 processors.
|
|
#
?
Mar 1, 2017 15:44
|
|
|
Pryor on Fire posted:drat this thread is really going off the reservation in the final days before ryzen. You guys it's gonna be ok, you don't have to talk about CPUs on the moon just hang on a little longer! Well, there's pretty much nothing else to talk about until tomorrow, all the leaks and whatnot are done, NDA lifts tomorrow, then the thread explodes.
|
|
#
?
Mar 1, 2017 15:51
|
|
|
SourKraut posted:What time is the embargo expected to drop? https://twitter.com/IanCutress/status/836541044581888000
|
|
#
?
Mar 1, 2017 15:57
|
|
|
Potato Salad posted:ISS uses 386. Hubble is on 486. Certification for flight takes a long rear end time. And it's worth remembering that the laptops and such which they use for the research are just normal modern Thinkpads most of the time. They've got about 50 or so thinkpads across the station.  FaustianQ posted:Why not an ARM solution? Or is legacy code holding them back? God NASA needs an actual loving budget. ARM solution for what? The air processing control and altitude stability controls etc already work, there's no point in replacing them. Perhaps when another module gets added onto the station they might use some ARM processors in the systems for that, but they will still need to be able to communicate with the rest of the station.
|
|
#
?
Mar 1, 2017 16:01
|
|
|
|
| # ? May 21, 2024 22:57 |
|
|
Didnt realize that Ubisoft purchased AMD
|
|
#
?
Mar 1, 2017 16:04
|
|