|
Tiny Tubesteak Tom posted:lol, the manual doesn't. It lists M.2SSD_1 AND M.2SSD_2 implying that there are multiple slots, but doesn't say anything else. It shows you a diagram of how to install a M.2 drive if you have no idea, but it has absolutely nothing about multiple slots or even that there is a difference between slots. The rest of the manual is just different languages Llamadeus fucked around with this message at 22:32 on May 23, 2019 |
# ? May 23, 2019 22:29 |
|
|
# ? May 17, 2024 01:48 |
|
Sometimes, there’s a separate layout leaflet.
|
# ? May 23, 2019 22:31 |
|
Llamadeus posted:If it's the X470 Taichi then page 8 has a literal picture of the motherboard with M2_1 and M2_2 labeled It is the Taichi, but page 8 for me is part of the list of specifications. Misprint confirmed? EDIT: I cannot install Windows whatsoever on this piece of poo poo. I get blue screens with different error codes every time the computer restarts. This has been the absolute worst nightmare of a build I have ever experienced. So much for the excitement to move up to X470. I have gotten as far as the first OOBE restart, but then suddenly it's "Windows could not configure the operating system for your hardware" and it just ends, lol. I am using the same USB drive with the same version of 1903 I just used to install the OS less than a week ago. Anyone ever heard of a CACHE_MANAGER blue screen? How about a 0x000021A? 0xc0000221? All of the documentation available online asks me to run troubleshooters that require you to be able to boot into an OS lmao. Then I get "Your PC/device needs to be repaired". Losing my mind barnold fucked around with this message at 22:58 on May 23, 2019 |
# ? May 23, 2019 22:33 |
|
list every component you are working with imho. i've seen stuff like this happen with hardware incompatibilities also maybe try a different HD/reseat the m2 E: VVVVVVVVVVVVVV i PM'd you a question so we dont derail this joint Worf fucked around with this message at 23:19 on May 23, 2019 |
# ? May 23, 2019 23:00 |
|
Ryzen 7 1700X ASRock X470 Taichi 4 x 8GB HyperX Fury DDR4-2400 RAM GIGABYTE Nvidia GTX 1070 WD Blue 500GB M.2 SATA SSD WD Black 2TB 7200RPM HDD Seagate Barricuda 1TB 7200RPM HDD Corsair RM850x PSU (brand new) that's it, that's all the components. right now I have taken every stick of RAM out except for one and we'll see what happens. I haven't been this stumped about computer issues in my entire life. The whole saga of diagnosing my B350 mobo and now problems extending into a completely new build is ultra lame. FAKE EDIT: While I was writing this post, Windows restarted into the OOBE on one stick of RAM. I don't get what's going on here. I have Schroedinger's RAM. Both the old kit I replaced and the new sticks are on the QVL.
|
# ? May 23, 2019 23:13 |
|
https://forums.tomshardware.com/threads/cannot-install-windows-10-on-new-ryzen-build.2979394/ that guy has some very similar stuff going on. im reading through it right now
|
# ? May 23, 2019 23:33 |
|
Statutory Ape posted:My AMD build is in this : Mines in a Streacom DA2. Rails attach to runners inside the case, radiators / fans can be mounted on them (anywhere, if it fits inside). You can invert the mobo layout too. I have RGB on the GPU, need to install an app to switch it off. I'm just happy it passive idles.
|
# ? May 24, 2019 00:03 |
|
GRINDCORE MEGGIDO posted:Mines in a Streacom DA2. Rails attach to runners inside the case, radiators / fans can be mounted on them (anywhere, if it fits inside). You can invert the mobo layout too. Whoa, that's pretty rad. I was at my local MicroCenter today looking at some of these nifty form factor cases. It's tough for me to get an idea of scale - is that a mATX case? There were a couple Lenovo AIO computers that were about the length and width of a CD jewel case. I haven't really looked at the prebuilt market in a long time so it was pretty cool to see that the whole "nettop" design actually went somewhere
|
# ? May 24, 2019 00:07 |
|
It's a SFF ITX case. It's ~17 liters, which puts it about halfway in between tiny SFF cases like the Dan Case A4 (7.5 liters) or NCase M1 (12.5 liters) and standard ITX mini towers like the NZXT H200 (26 liters) or Fractal Design Nano S (27 liters). Here's the DA2 (right) next to an NCase M1 (middle) and an M1 (left) vs. a Nano S (right) E: Okay, Imgur bbcode apparently isn't working for me after their site update. What am I doing wrong? Here's the link to the pictures. E2: It no longer automatically adds ".jpg" to the end of the link in bbcode - need to do that yourself, apparently. Stickman fucked around with this message at 00:36 on May 24, 2019 |
# ? May 24, 2019 00:30 |
|
There's a 2700x in there, undervolted. I was running a fury X with it for a while, did surprisingly well. I'll replace the 2060 in there with a Navi probably. Just hope it's not a space heater. Nvidia drivers are like going back 15 years. GRINDCORE MEGGIDO fucked around with this message at 01:06 on May 24, 2019 |
# ? May 24, 2019 01:02 |
|
https://www.anandtech.com/show/14380/sugon-workstation-with-hugon-dhyana-cpu Bad for Intel? Good for AMD? Great for China?
|
# ? May 24, 2019 01:54 |
|
LRADIKAL posted:https://www.anandtech.com/show/14380/sugon-workstation-with-hugon-dhyana-cpu If they get released in the US, but I doubt they'll gain much market share. The return of VIA?
|
# ? May 24, 2019 02:56 |
|
?? VIA is Taiwanese you dummy!
|
# ? May 24, 2019 03:30 |
|
CommieGIR posted:If they get released in the US, but I doubt they'll gain much market share. Can't be sold outside of China due to the x86 license deal with Intel, I don't think. I really do hope someone in the west gets their hands on one, it sounds like its exactly a Zen1 CPU with some China specific security features. Seamonster posted:?? VIA is Taiwanese you dummy! I think hes referring to VIA being the 3rd x86 license holder, not that they were Chinese.
|
# ? May 24, 2019 03:38 |
|
Cygni posted:Can't be sold outside of China due to the x86 license deal with Intel, I don't think. I really do hope someone in the west gets their hands on one, it sounds like its exactly a Zen1 CPU with some China specific security features. "Security Features". I wonder what a forensic look at the BMC/SOC code would reveal, given the nature of the remote management tools they seem to all have these days.
|
# ? May 24, 2019 03:41 |
|
Seamonster posted:?? VIA is Taiwanese you dummy! He got it, but I meant x86 market share, not base of manufacturing.
|
# ? May 24, 2019 04:04 |
|
CommieGIR posted:If they get released in the US, but I doubt they'll gain much market share. The US? These are for China, which is probably the largest market for processors in the world!
|
# ? May 24, 2019 04:59 |
|
Combat Pretzel posted:Navi is supposedly going double the width. What is that expected to do? Will it get close to anywhere near the 2080s? That rumor is a bit suspect since apparently the commits in llvm for Navi still use 64 as the wavefront size.
|
# ? May 24, 2019 12:12 |
|
Tiny Tubesteak Tom posted:FAKE EDIT: While I was writing this post, Windows restarted into the OOBE on one stick of RAM. I don't get what's going on here. I have Schroedinger's RAM. Both the old kit I replaced and the new sticks are on the QVL. I'd setup two sticks in proper dual channel, run memtest briefly, remove them and test the other two. Then you know your RAM is probably good. From there, put all 4 in and try the ryzen dram calculator and use those settings. I need more information about your RAM to tell you which memory type you should pick, but memory rank 1, dimm modules 4, and HyperX uses a bunch of different Hynix or Micron dies so you need the exact model number to figure out which yours uses. Khorne fucked around with this message at 15:14 on May 24, 2019 |
# ? May 24, 2019 14:09 |
|
Do the Spectre/Meltdown/etc. mitigations handicap Sandy Bridge/Ivy Bridge enough to make Bulldozer a worthwhile purchase in hindsight? Phoronix tested this and no, of course not. They still suck rear end; the only Intel chip any of the FX CPUs manages to beat by a significant margin is an i3-2120 with HT disabled.
|
# ? May 24, 2019 15:31 |
|
Mr.Radar posted:Do the Spectre/Meltdown/etc. mitigations handicap Sandy Bridge/Ivy Bridge enough to make Bulldozer a worthwhile purchase in hindsight? Phoronix tested this and no, of course not. They still suck rear end; the only Intel chip any of the FX CPUs manages to beat by a significant margin is an i3-2120 with HT disabled. Holy crap hahahaha. I have lived with an FX-6300 CPU from 2012 to 2018, wonder how I managed it. (BTW still going strong as a living room PC connected to the TV, an FX 6300 + 8GB ram is still more than good enough for internet, movies and some image editing)
|
# ? May 24, 2019 15:37 |
|
Khorne posted:4 stick on ryzen gen1 probably requires lower clocks or tinkering with timings and voltage. Gen1 in particular struggles with 4 sticks. Had a hunch this might be the case. In 2017 I built this rig originally knowing that Ryzen had some latent fuckery with RAM compatibility. I had great success with the initial build and never had a problem until moving to four sticks. I'm prepared to move up once the next generation CPUs are on the market, hence the inclination to go to the X470 board now. Right now I have it running fairly stable with A-XMP enabled, 2400 manually set as the frequency, AND the timings loosened to 16-16-16-36. I have determined that I can boot into the BIOS no matter what the RAM configuration is on the mobo, which is a big step up from the beeps of death I was getting on the B350, and it's only Windows that starts sputtering when things are wonky. barnold fucked around with this message at 18:42 on May 24, 2019 |
# ? May 24, 2019 18:39 |
|
I gotta say, it's super nice to be able to stream games and not having to use my capture card.
|
# ? May 25, 2019 13:37 |
|
It's like having two PCs in one!
|
# ? May 25, 2019 14:23 |
If you're on a newer Nvidia card you can always stream with NVENC which is decent quality for minimal cpu hit
|
|
# ? May 25, 2019 14:32 |
|
SSJ_naruto_2003 posted:If you're on a newer Nvidia card you can always stream with NVENC which is decent quality for minimal cpu hit But my frames
|
# ? May 25, 2019 16:22 |
|
Deuce posted:But my frames Generally NVENC loses less frames than CPU streaming would if your CPU is too weak for CPU streaming. 'course the downside is you need way more bitrate to make it look as good as cpu streaming (probably too much to stream it unless you're on at least 30 mbit up). NVENC has been improving each nvidia generation though. MagusDraco fucked around with this message at 20:48 on May 25, 2019 |
# ? May 25, 2019 16:58 |
|
Clock for clock, a single Matisse core matches a Coffee Lake core in Geekbench 4. There's for sure some variance in the performance per task, not sure what that means for real world usage. While the test page says one was at 2.89GHz and the other at 3.2GHz, the data for both tests shows that the median clocks where 3.975GHz and 3.983GHz respectively.
|
# ? May 25, 2019 20:00 |
|
|
# ? May 25, 2019 20:22 |
|
Please let this be true in practical workloads, please
|
# ? May 25, 2019 21:20 |
|
Maybe the gimpy garbage ram on the Intel platform has a lot to do with it.
|
# ? May 25, 2019 21:40 |
|
K8.0 posted:Maybe the gimpy garbage ram on the Intel platform has a lot to do with it. LMAO yeah, 2133MHz vs 2666MHz isn't apples to apples. And even then the AMD part loses in a lot. And has higher mem latency (as expected due to IO die) No contest in multi-core due to HT
|
# ? May 25, 2019 21:48 |
|
Malcolm XML posted:No contest in multi-core due to HT "Please turn HT off, because its a security risk" - Intel
|
# ? May 25, 2019 21:52 |
|
Anybody want to dig up a 2600 comparison at the same frequency. I tried to do it but phonepostin
|
# ? May 25, 2019 21:59 |
|
CommieGIR posted:"Please turn HT off, because its a security risk" - Intel
|
# ? May 25, 2019 22:13 |
|
Cygni posted:Anybody want to dig up a 2600 comparison at the same frequency. I tried to do it but phonepostin WCCFTech, so consult your practitioner on sodium intake: https://wccftech.com/amd-zen-2-6-core-cpu-benchmark-leaked-faster-than-ryzen-7-2700x-in-geekbench-4/ WCCFTech posted:[This sample] managed to achieve a single-core score of 5061 and a multi-core score of 25481. In contrast, a Ryzen 7 2700X scores 4923 points and 25209 points respectively
|
# ? May 26, 2019 01:41 |
|
NewFatMike posted:WCCFTech, so consult your practitioner on sodium intake: 8700k scores ~5880 single core and ~26050 multicore. At 5GHz, ~6500 and ~28800 without any of the exploit mitigations. So 5061 and ~25000 aren't that impressive. It's impressive in the sense that if it is actually locked to 4GHz it will out perform a 4GHz 8700k by ~4%-5% in geekbench. It's impressive in the sense that it will probably cost much less. It's slightly worrying because a 25% clock speed difference between a 3300 and the rumored 5GHz boost CPUs is somewhat odd given the total lack of variance in previous iterations of zen. They are on an entirely different foundry's entirely different node whose consistency should be suffering from the lack of EUV. So maybe it's possible. Tomorrow night at 10pm est we'll hopefully see! Khorne fucked around with this message at 01:59 on May 26, 2019 |
# ? May 26, 2019 01:46 |
|
Measly Twerp posted:Clock for clock, a single Matisse core matches a Coffee Lake core in Geekbench 4. There's for sure some variance in the performance per task, not sure what that means for real world usage. Holy poo poo that is impressive. The lowest end Ryzen 3 is edging out their previous highest-end 2700X, and doing that with two less cores. It also confirms AdoredTV's leaks, which most people thought was too good to be true horseshit. As a bonus vs Intel, the security model of AMD isn't based on Swiss Cheese and you don't have to disable any critical, market-segmenting features. And Intel has no response for at least two years. 10nm will be laptop-only (meaning it was a bust), and nothing to fall back on like their Pentium 3-M when Prescott failed.
|
# ? May 26, 2019 06:10 |
|
https://hothardware.com/news/amd-ryzen-3-3300-zen-2-6-core-12-thread-cpu-benchmarks Is real?
|
# ? May 26, 2019 09:10 |
|
|
# ? May 17, 2024 01:48 |
|
That's a big fat maybe. It's now just over 12 hours until Dr. Su's Computex keynote where we'll presumably get some real info.
|
# ? May 26, 2019 14:28 |