|
Craptacular! posted:Given that I'm looking at a Raven Ridge APU, I guess I should assume none of the motherboards available now will work with it out of the box? They may need a BIOS update. And besides, all it takes is one joke that doesn't come across as a joke, that's escalated into a personal attack, and we'll have goons clawing at eachother's throats in no time. Anime Schoolgirl posted:it would actually legitimately own just not in a productive way SwissArmyDruid fucked around with this message at 23:25 on Jan 12, 2018 |
# ? Jan 12, 2018 23:23 |
|
|
# ? May 17, 2024 07:58 |
|
I suppose. This forum seems to be grown up and mature to not fall into feeling the need to justify our purchases and tear down other peoples. Lots of people cross-post and the one time someone cast shade at Paul a few days ago felt like the first time in months I'd seen something like that between the Intel, AMD, and GPU threads and threatened to damage the generally good camaraderie. Anyway, nothing I can do but wait and see what happens. Apparently ASRock is pretty good at POSTing with CPUs that haven't been produced yet, reports are when Ryzen 5 launched they were good enough to reach UEFI and flash an upgrade, whereas Asus and MSI was not. As this is a xXxTEAM BLOOxXx HAUS there is no way I'd be able to flash if it didn't work. Here's hoping.
|
# ? Jan 12, 2018 23:33 |
|
....I gotta say, though. If I were in your situation, I would grab an R3 1200 on sale and keep that around purely for bootstrapping all motherboards going forward. I mean, at $109 retail with the new pricing, you absolutely could afford to, but I'm sure that somebody has it on sale for under a c-note. I'm sure you don't need the reassurance, but ASRock is where I would be going in terms of boards in your situation too, so I think you made the right choice. SwissArmyDruid fucked around with this message at 23:58 on Jan 12, 2018 |
# ? Jan 12, 2018 23:53 |
|
Yeah, no. I'm looking at moving one computer that's a basic file server and x264 encoder to that $99 APU and that's probably it, and it's intentionally being done as low-cost as possible because going over budget will eat rent and other basic survival needs. I mean, technically speaking, the Haswell i3 that's on there is doing the job just fine, I'd just like to take it's DDR3 RAM for my primary and this seems like a better purchase than buying more obsolete RAM second-hand.
|
# ? Jan 12, 2018 23:58 |
|
Oh, my bad. I incorrectly inferred that this was a work situation with someone else's money paying for it. Yeah, I'd still have gone with the ASRock board anyways.
|
# ? Jan 13, 2018 00:02 |
|
I'd just go with ASRock as a general rule for boards because I've never had a bad experience with them like say ASUS or Gigabyte and support seems pretty solid.
|
# ? Jan 13, 2018 00:23 |
|
So if I planned on doing a system upgrade to TR4 in November, should I just wait for Zen 2 in 2019? Moving away from 4790k which just can't keep up with encoding for streaming and playing. Planned on going max core count on the CPU with an Enermax TR4 AIO unit for cooling, 32GB of through 4x8GB somewhere in the neighborhood of 3600+ rated. Board choice would be either ASRock or MSI with built-in wifi.
|
# ? Jan 13, 2018 00:37 |
|
Wouldn't it be more cost efficient at that level to just use the 4790K as a stream-dedicated PC and buy a Ryzen+ to play with? I bought my current rig five years ago with the intention of doing both, wound up streaming a few dozen hours and getting five followers and realizing that this really isn't going to replace A Real Job so I shouldn't invest high end hardware into it. If you want to stop using GPU encoding, just retire the old PC to capturing and encoding with OBS's NDI plugin.
|
# ? Jan 13, 2018 00:49 |
|
Craptacular! posted:I can't see it? Intel users seem to be pretty self-aware of the company's failings. Craptacular! posted:I've spent most of the past 24 hours trying to find out if the budget Ryzen motherboards are stable and finding that nobody in charge of having tech opinions actually buys them. If you're buying for home/enthusiast use so long as it has a BIOS update with the latest public AGESA version almost all of them are perfectly fine. If you're overclocking you'll want one with a good VRM and maaaaybe a external clock gen if you want to try and max it as much as possible but the latter is usually not really needed especially if you're sticking to air cooling with sensible (ie. 1.3-1.4v tops) overvolting. SlayVus posted:So if I planned on doing a system upgrade to TR4 in November, should I just wait for Zen 2 in 2019? At this point its looking to be a few months difference of wait and that sort of wait is worth it if the jump is even half as big as its rumored to be. PC LOAD LETTER fucked around with this message at 01:53 on Jan 13, 2018 |
# ? Jan 13, 2018 01:33 |
|
Craptacular! posted:I suppose. This forum seems to be grown up and mature to not fall into feeling the need to justify our purchases and tear down other peoples. Lots of people cross-post and the one time someone cast shade at Paul a few days ago felt like the first time in months I'd seen something like that between the Intel, AMD, and GPU threads and threatened to damage the generally good camaraderie. Eh. Yeah. Sorry about that. I didn't set out to have a go at Paul. I posted an AdoredTV video and that guy just can't help use inflamitory titles for his videos, which is a shame because it was actually a pretty good video that had a point other than "lol Intel fail" which is what you'd expect from the title. Then Paul chimed in and I knew he hadn't watched it, so I got annoyed and shatpost instead of explaining why it was probably a good video to watch.
|
# ? Jan 13, 2018 02:39 |
|
PC LOAD LETTER posted:At this point its looking to be a few months difference of wait and that sort of wait is worth it if the jump is even half as big as its rumored to be.
|
# ? Jan 13, 2018 03:29 |
|
Measly Twerp posted:Eh. Yeah. Sorry about that. I didn't set out to have a go at Paul. I posted an AdoredTV video and that guy just can't help use inflamitory titles for his videos, which is a shame because it was actually a pretty good video that had a point other than "lol Intel fail" which is what you'd expect from the title. Uh, I watched it, and I didn't see much different than Paul. He found an 8700 in an OEM box that was thermal throttling, and then lifted all sorts of speculations about the quality of the 8400. There's evidence in the form of people who actually have ordered these chips that they do work properly when they aren't slapped in by a careless vendor.
|
# ? Jan 13, 2018 03:49 |
|
Combat Pretzel posted:Depends how you define a couple of months. With Zen+, TR+ seems like at least 3 months behind, with what that super precise H2/18 timeline. They will still essentially be the same die AFAIK, the difference will be in the package/validation which is the whole advantage with AMD's approach.
|
# ? Jan 13, 2018 04:53 |
|
Craptacular! posted:Uh, I watched it, and I didn't see much different than Paul. He found an 8700 in an OEM box that was thermal throttling, and then lifted all sorts of speculations about the quality of the 8400. There's evidence in the form of people who actually have ordered these chips that they do work properly when they aren't slapped in by a careless vendor. Comically, I have an 8700 and 8400 and skipped watching that video because I can't take his accent. He's not wrong. At base and completely stock, the 8700 and 8400 will boost up to the max p-state (that intel doesn't guarantee) for a few seconds, but then crank back cause of TDP restriction in the motherboard. Nothing to do with cooling. I assume thats why Intel stopped guaranteeing anything other than 1 core boost clocks. 6 cores at 1.264v, 4.3ghz is just not gonna happen in 65w TDP sustained. So on that part, he's right. But if you just go in the motherboard (all the retail boards I've tried have it) and change the TDP to 95watts or higher, it will run at the max in the p-state tables forever. I've never had any stability problems at all from either.. Should also note that both the Asus and Gigabyte boards I've used didn't do it automatically, you had to manually set it. Also, MCE does nothing for either chip. That said, if you get a PC in a box with an 8700 or 8400, your mileage may vary with what the BIOS offers. So any dumb "low quality" silicon or "high endmotherboard manufacturers are CHEATING" claims hes hucking is his normal dumb AMD fan-service conspiracy theory crap. if you consider that a conjob, well, knock yourself out i guess.
|
# ? Jan 13, 2018 05:46 |
|
What I took away from it is that reviewers need to do a better job explaining these things, and even mitigating against them when possible or it's going to become very difficult to compare things going forward when one 8700 is inexplicably performing worse than others. And that we'll also see AMD processors do this as soon as they have a process with more headroom. It's not going to be possible to say "this computer has X processor, therefore ..." in the future without caveats. Also his accent is extremely light, much easier than a Newfoundland or Wisconsin accent.
|
# ? Jan 13, 2018 06:36 |
|
I have no issues with his accent, partially because he does a good job of highlighting segments of his web pages as he reads them. Now, on the other hand, the Welsh...Cygni posted:At base and completely stock, the 8700 and 8400 will boost up to the max p-state (that intel doesn't guarantee) for a few seconds, but then crank back cause of TDP restriction in the motherboard. Nothing to do with cooling. I assume thats why Intel stopped guaranteeing anything other than 1 core boost clocks. 6 cores at 1.264v, 4.3ghz is just not gonna happen in 65w TDP sustained. So on that part, he's right. Well, sure, turbo is turbo, and the 3.2 base clock and 4.6 turbo is quite a wide gap. I have seen one reviewer I trust to be straightforward say he got "4.3ghz all day" out of an 8700, but also he had high-end boards (MSI Godlike and Gigabyte Aorus) to test on.
|
# ? Jan 13, 2018 07:14 |
|
The possibility of APUB mining has come up here. I don't know anything that technical about it, but is it possible increased latency from using system RAM might kill mining appeal?
|
# ? Jan 14, 2018 21:32 |
|
Lenovo 720S ultrabook on sale now in some parts of the world with an AMD APU, 15W TDP. https://www.youtube.com/watch?v=t3YsZT0ps4U
|
# ? Jan 16, 2018 02:50 |
|
I think that’s the one that only has single channel DDR4 2400. I’ll be really curious to see if it does better than the x360 with dual channel RAM and a goosed R5 2500U at 25W.
|
# ? Jan 16, 2018 04:03 |
|
It had better not be single-channel! R7 APU and only single-channel, what the hell.
|
# ? Jan 16, 2018 04:14 |
|
Oh man I was wrong. It's single channel 2133
|
# ? Jan 16, 2018 04:19 |
|
Yeah, and we saw that with the HP, but then HP decided to not be stupid and did dual-channel after all.
|
# ? Jan 16, 2018 04:26 |
|
At the reveal event the HP was originally posted with dual channel? Lenovo were the only ones with single channel, I believe:
|
# ? Jan 16, 2018 04:33 |
|
Shame it's gimped, quite a nice looking machine.
|
# ? Jan 16, 2018 04:37 |
|
I’m hoping the KBL-G and Raven Ridge in Not poo poo™️ SKUs will cause some increased momentum with OEMs making systems for everything. Hopefully some enterprise class laptops with decent cooling will crop up sooner than later. Folks who game are getting older, and it would be nice not to have to choose between work and play computers. And also GPU prices are still astronomic.
|
# ? Jan 16, 2018 05:17 |
|
How much does single channel and dual channel matter? GamersNexus did some tests with video games and found the difference varied depending on the game but usually was not very much. I'm looking at building a cheapo Ryzen system with a single stick of RAM for now to compensate for the fact that the price is so high.
|
# ? Jan 16, 2018 05:20 |
For discrete, yeah its not that important since they have their own memory domain but for integrated/cpu package GPUs they share the system RAM, which has a higher latency.
|
|
# ? Jan 16, 2018 05:37 |
|
NewFatMike posted:At the reveal event the HP was originally posted with dual channel? Lenovo were the only ones with single channel, I believe: Then I have to blame my faulty own lovely single-channel non-ECC memory, because I swear I remembered that the HP and Lenovo were both cruising for soldered single-channel.
|
# ? Jan 16, 2018 05:51 |
|
Craptacular! posted:How much does single channel and dual channel matter? For CPU only performance the difference will be a different story, it won't be anywhere near as bad, but you'll still see some sort of a performance hit. If you have a dGPU you'll only have to worry about the CPU performance hit but yeah it won't be too terrible.
|
# ? Jan 16, 2018 06:03 |
|
SwissArmyDruid posted:Then I have to blame my faulty own lovely single-channel non-ECC memory, because I swear I remembered that the HP and Lenovo were both cruising for soldered single-channel. It's okay, I still think you're a good poster and you contribute good things to the hardware threads.
|
# ? Jan 16, 2018 06:07 |
|
PC LOAD LETTER posted:For iGPU performance its pretty much a deal breaker. Like a 40-50% performance hit. Possibly worse at times. Will this matter if it's a Linux based media player and bittorrent box? I just want to run a Plex server that can trascode because my pokey ancient ARM-based NAS device can't. There won't be any gaming happening on the actual silicon (I'd do Steam in-home streaming from my i7 dGPU if I was into that.)
|
# ? Jan 16, 2018 06:16 |
|
Craptacular! posted:Will this matter if it's a Linux based media player and bittorrent box? I just want to run a Plex server that can trascode because my pokey ancient ARM-based NAS device can't. There won't be any gaming happening on the actual silicon (I'd do Steam in-home streaming from my i7 dGPU if I was into that.) Sorry
|
# ? Jan 16, 2018 06:18 |
|
On some Intel iGPUs you could get a noticeable performance increase by using double ranked RAM. You could have more DRAM pages* open or something. *: The DIMM page kind, not the x86 page kind.
|
# ? Jan 16, 2018 06:20 |
|
PC LOAD LETTER posted:I want to say it won't be a big deal, that stuff is either mostly CPU dependent That's my hope. Going from 2c/4t/Win10/8GB DDR3 to 4c/4t/Linux/4GB DDR4 for quick and dirty h264 encodes and not suffering noticeably until RAM prices fall.
|
# ? Jan 16, 2018 06:29 |
|
Intel acknowledges Haswell and Broadwell reboots after patches AMD will issue optional Ryzen and Epyc microcode updates for Spectre IBM POWER Confirmed Impacted by Security Design Flaws I'm begging you, AMD. Please turn this all into a win. Bring balance back to the SwissArmyDruid fucked around with this message at 11:10 on Jan 16, 2018 |
# ? Jan 16, 2018 10:51 |
|
Have they been working on the patch since last June
|
# ? Jan 16, 2018 11:47 |
|
Who, Intel? Probably. A thought occurred to me as I drank my morning tea, though. (Yes, I do occasionally have those. Good ones, even. Thoughts, I mean, not tea. I have that every morning.) The road ahead for AMD is not going to be easy. There is a very real risk that even despite all of..... <gesticulates nebulously> "this" around Spectre and Meltdown? That this continues to be a banner year for Intel, by virtue of any perf hits forcing people to move off of the same old Sandy/Ivy/Haswell/Broadwell machines they've been sitting on because of a complete lack of need, to Kaby Lake/Cannon Lake, because it will require less effort to just drop in Intel's latest, rather than go through the process of finding what AMD products suit the current need. Because as everybody in IT knows: the most powerful force in the world is inertia. Basically, Intel could wind up inadvertantly pulling a Microsoft moving everyone to Win 10 and make some very happy shareholders... as well as one very unhappy Mr. Brian Krzanich Esq. for panic dumping his stock not only illegally, but also at a loss. SwissArmyDruid fucked around with this message at 17:01 on Jan 16, 2018 |
# ? Jan 16, 2018 15:51 |
|
I think Intel will make out like bandits. But it'll be interesting to see what happens if some other flaw gets as well publicised as meltdown, after people replace old Intel chips with coffeelake. But who knows it might be AMD being hit with something. GRINDCORE MEGGIDO fucked around with this message at 16:35 on Jan 16, 2018 |
# ? Jan 16, 2018 16:29 |
|
You do raise a good point, although annual socket changes and a change from DDR3 to DDR4 over the course of the most affected Intel products probably shaves off a fair number of users. If I were AMD, I would push the long-term socket support advantage, not necessarily a security one. Even if Meltdown or a similar vector impacts AMD platforms you're only out a processor instead of processor+Mobo and maybe RAM. Although that benefit if limited by Zen+ being around the corner.
|
# ? Jan 16, 2018 16:29 |
|
|
# ? May 17, 2024 07:58 |
|
I'm not going to buy a new computer till I come back from mob which is will be in like a year and a quarter so by then I'll decide if an AMD processor is for me. Zen already sort of spooked Intel to release 6 core processors so hopefully their new offerings are even more competitive
|
# ? Jan 16, 2018 16:37 |