|
Eletriarnation posted:I'd assume not. I considered it for a while first because it also has an OK VRM setup, dual M.2, ECC support, and three full-length card slots which were some of the things I wanted but I ended up going with a Taichi for a few more bells and whistles. It would be awesome if ASRock does something similar since that's the only downside of X470 I can see myself potentially caring about. It's not the tweet, twitter itself is absolutely fuckered.
|
#
?
Jul 11, 2019 20:16
|
|
|
|
| # ? May 31, 2024 22:31 |
|
|
Eletriarnation posted:Sorry, which conclusion do you have a problem with? That the mobo is magically pulling the extra 50W and not the CPU itself. That this is somehow the chipset's fault and not mobo vendors being dumb with stock settings. That this is somehow a new discovery and not the ol' ASUS BIOS switcheroo part 2. Arzachel fucked around with this message at 20:28 on Jul 11, 2019 |
|
#
?
Jul 11, 2019 20:22
|
|
|
It's because Twitter is down, I imagine the tweet is still up.
|
|
#
?
Jul 11, 2019 20:27
|
|
|
Arzachel posted:That the mobo is magically pulling the extra 50W and not the CPU itself. That this is somehow the chipset's fault and not mobo vendors being dumb with stock settings. That this is somehow a new discovery and not the ol' ASUS BIOS switcheroo part 2. OK, well, you aren't the guy I was asking but I don't think I said any of those things. What I said is that if you run the same chip on an X570 board as on an X470 board doing the same tasks, the whole setup seems to use more power. (Didn't even say 50W, and I care more about idle power than load anyway.) I do not care if the chipset is using that power, or if some other component on the motherboard is using it, or if it's actually something that the motherboard is causing the CPU to do - the salient difference is that more power is being used, and I as the end user will have to pay my power company for it. The bit about this being not "somehow a new discovery" makes no sense to me. This chipset came out four days ago and these are pretty much the only numbers I have seen at this point; there are very few articles comparing X470 to X570 apples-to-apples yet. Do you have something older, or are you talking about something different that happened with a different platform and saying that it's relevant to this situation, or am I misunderstanding you entirely? Eletriarnation fucked around with this message at 20:52 on Jul 11, 2019 |
|
#
?
Jul 11, 2019 20:49
|
|
|
Regardless of X570 using more power (but probably not significantly more), mainboard vendors pre-configuring their BIOS to use significantly more power by increasing boost clocks etc. to make their boards seem like better performers is something that has been happening for years and needs to be controlled for when comparing chipset power usage.
|
|
#
?
Jul 11, 2019 20:56
|
|
|
Mr.Radar posted:On the consumer OSes Microsoft is restricting Windows Updates only to "supported" processors (for Windows 7 and 8.1 that's CPUs released before Dec 31, 2016) unless you do some hacks to work around it. I'm not sure if that applies to server OSes as well. So server 2012 is paired with Windows 8.x. If Zen 2 is going to run on a server and does not support windows 8.x you will need server 2016 for windows 10 compatibility. Server 2019 would work as well. This is assuming you can get everything running. Some software just refuses to run on a server OS.
|
|
#
?
Jul 11, 2019 20:57
|
|
|
Lambert posted:Regardless of X570 using more power (but probably not significantly more), mainboard vendors pre-configuring their BIOS to use significantly more power by increasing boost clocks etc. to make their boards seem like better performers is something that has been happening for years and needs to be controlled for when comparing chipset power usage. This is probably the key part. The chipset itself doesn't seem to draw more than 10W, even when heavily utilizing PCIe 4.0.  I'm wondering if maybe the X470 boards in the Extremetech article were affected by the boost clock bug, and maybe idle clocks too? They didn't mention clock speeds at all. I think the two big take-aways of the article are 1) something funny is going on with either X570 overprovisioning or X470 underprovisioning (but not with the chipset power consumption itself), and 2) more reviewers need to start directly measuring the power consumption of individual components rather than relying on total system consumption. Stickman fucked around with this message at 21:38 on Jul 11, 2019 |
|
#
?
Jul 11, 2019 21:04
|
|
|
First impressions: the 3900X is a monster. I have days of WCG and undervolt/underclocking tests left to do, but it was quick to get results for Stockfish and OpenFOAM. https://firepear.net/grid/ryzen3900/#stockfish-chess-engine TLDR: Running stock, straight out of the box, no system differences but a CPU swap: 2.47X the performance of the 1600 on Stockfish; 1.7X on OpenFOAM. There's clocks and power details in the doc.
|
|
#
?
Jul 11, 2019 21:10
|
|
|
I've been trying to figure out if this X570 motherboard is t topology or daisy chain for RAM. Does anyone here know?
|
|
#
?
Jul 11, 2019 21:12
|
|
|
Eletriarnation posted:OK, well, you aren't the guy I was asking but I don't think I said any of those things. https://www.overclock.net/forum/10-amd-cpus/1728758-strictly-technical-matisse-not-really.html The interesting part: quote:The peak power difference between the faulty and the fixed bioses is around 35W (Prime95).
|
|
#
?
Jul 11, 2019 21:36
|
|
|
Zotix posted:The x570 taichi looks like a very solid board. What exactly is the difference between t topology vs daisy chain? The taichi uses t topology according to this spreadsheet that has been floating around for a week. Many apologies, dude, I forgot to answer your question and only just remembered when this video floated back to the top of my youtube suggestions. Buildzoid did a video that goes into depth into the difference between the two and their relative benefits. https://www.youtube.com/watch?v=3vQwGGbW1AE
|
|
#
?
Jul 11, 2019 21:38
|
|
|
That third party seller has more 3900X's in stock at MSRP. 7 as of this post. https://www.amazon.com/AMD-Ryzen-3900X-24-Thread-Processor/dp/B07SXMZLP9 Edit: And gone. 
8-bit Miniboss fucked around with this message at 21:47 on Jul 11, 2019 |
|
#
?
Jul 11, 2019 21:42
|
|
|
Soricidus posted:Thanks for the effort post! Greatly appreciated. Check Reddit for cheap e die sticks of ram that are recommended for overclocking. What temps are people getting Vs second Gen? Also ordered an nhd15s and dug my old evga supernova out of a spare case. It's going in one of these:  I don't think cooling will be a problem.
|
|
#
?
Jul 11, 2019 21:44
|
|
|
Anyone tested a 3700x in a fresh Asus 470 Prime Pro without needing a BIOS flash yet? If so what vendor did you get it from that has updated it already?
|
|
#
?
Jul 11, 2019 21:47
|
|
|
GRINDCORE MEGGIDO posted:Check Reddit for cheap e die sticks of ram that are recommended for overclocking. Open cases don't allow for good directional airflow and will generally end up with components running hotter and making more noise.
|
|
#
?
Jul 11, 2019 21:54
|
|
|
I didn't think they could make an uglier case than the ones with infinity mirrors, but here we are.
|
|
#
?
Jul 11, 2019 21:57
|
|
|
mdxi posted:it was quick to get results for Stockfish Thanks again!
|
|
#
?
Jul 11, 2019 22:04
|
|
|
BangersInMyKnickers posted:Open cases don't allow for good directional airflow and will generally end up with components running hotter and making more noise. I've run open frame cases, zero problems. It's a test bench with a window on the side. taqueso posted:I didn't think they could make an uglier case than the ones with infinity mirrors, but here we are. GRINDCORE MEGGIDO fucked around with this message at 22:24 on Jul 11, 2019 |
|
#
?
Jul 11, 2019 22:21
|
|
|
taqueso posted:I didn't think they could make an uglier case than the ones with infinity mirrors, but here we are. I dunno I unironically kind of like the "Honda gasoline generator" vibe.
|
|
#
?
Jul 11, 2019 22:22
|
|
|
GRINDCORE MEGGIDO posted:I've run open frame cases, zero problems. It's a test bench with a window on the side. I didn't say it wouldn't work, I said it would be louder and hotter. Heatsinks need cfm to work, and having shrouding and directional airflow ensures that a consistent airflow of room temp air in moving in to the case to collect heat and be pushed out the back, and the shrouding will deflect/absorb sound. With an open case, you're getting whatever air happens to drift in there and it isn't moving nearly as much so it has more time to collect heat from other components before it moves through whatever important heatsink, then the heat gradually leaves via convection if it doesn't get pulled back around to the inlet side of the same heatsink it just left.
|
|
#
?
Jul 11, 2019 22:28
|
|
|
I didn't say you said it wouldn't work (?) It depends what you want to run. In my test bench, I get better ambient, and better loaded temps then any case config I've tried, except with some fan noise (or just fans, I don't need them, temps are fine). ... What temps are people getting with a 3 series chip? GRINDCORE MEGGIDO fucked around with this message at 23:02 on Jul 11, 2019 |
|
#
?
Jul 11, 2019 22:34
|
|
|
GRINDCORE MEGGIDO posted:I like it's not built of garbage idk it looks like someone repurposed some go-kart wreckage
|
|
#
?
Jul 11, 2019 22:41
|
|
|
Inept posted:idk it looks like someone repurposed some go-kart wreckage I'll keep it in a different room to the Lian lis.
|
|
#
?
Jul 11, 2019 22:44
|
|
|
Welp, I knew something was going to go wrong. Looks like this board actually needs the next BIOS version. I ordered a cheapo CPU from Amazon so I can flash the BIOS and then I'll either just return it or sell it on craigslist.
|
|
#
?
Jul 11, 2019 23:14
|
|
|
You know you can get a free proc from AMD and return it after the reflash, right? Searching for the link now. e: Bottom of this page. e2: weird, they don't care if you return the HSF? Must be an extremely cheap model. e3: and they want proof from you that you contacted the mobo manufacturer for help first and were denied. Never mind, this is pretty disappointing. Eletriarnation fucked around with this message at 23:29 on Jul 11, 2019 |
|
#
?
Jul 11, 2019 23:22
|
|
|
Cojawfee posted:Welp, I knew something was going to go wrong. Looks like this board actually needs the next BIOS version. I ordered a cheapo CPU from Amazon so I can flash the BIOS and then I'll either just return it or sell it on craigslist. Confirmed my fears with the 470 Pro. AMD is offering kits that allow you to upgrade but I don't know the timeframe on them, I ended up getting a B450 Tomahawk just because I wasn't sure and didn't want to end up in a similar boat but sucks to hear it confirmed 
|
|
#
?
Jul 11, 2019 23:23
|
|
|
Eletriarnation posted:You know you can get a free proc from AMD and return it after the reflash, right? Searching for the link now. That will take days when I could just do it tomorrow.
|
|
#
?
Jul 11, 2019 23:27
|
|
|
Yeah, as the edits indicate my opinion of the offer diminished pretty dramatically as I read through the fine print. I might end up doing the same and building a HTPC with the 200GE, since even if ASRock is willing to give me an RMA over the EFI version I'm not sure if I'll want to wait on them.
|
|
#
?
Jul 11, 2019 23:31
|
|
|
Yeah that deal is absolutely insulting to anyone trying to build a new system, I get that it's probably more the board manufacturer's issue but wow what a ridiculous set of criteria just to try and get my new flagship AMD machine work. Thanks a lot for buying our product, hope you already have one or you're hosed outside a few specific configs! It does say they expect the kit returned within 10 days also, not sure where you were seeing that they don't want it back.
|
|
#
?
Jul 11, 2019 23:36
|
|
|
Eletriarnation posted:e3: and they want proof from you that you contacted the mobo manufacturer for help first and were denied. Never mind, this is pretty disappointing. lol AMD is trying to send a message to OEMs. "look if you insist on not doing flashback and you factory flash a BIOS version from 2017 then we're going to blow up your support queue, have fun"
|
|
#
?
Jul 11, 2019 23:36
|
|
|
Scorps posted:It does say they expect the kit returned within 10 days also, not sure where you were seeing that they don't want it back. They want the CPU back. They don't want the thermal solution/heatsink back.
|
|
#
?
Jul 12, 2019 00:31
|
|
|
fyi, the CPUs they used to send out were comedy option dead stock Bristol Ridge A6-9500's, and AMD was destroying them when you shipped them back. They make you ship it back so you have to actually go get a Zen CPU and don't just keep the terrible A6 and use that as the system. Giving someone a computer with an A6-9500 is a crime against humanity, so its probably for the best they are getting junked.
|
|
#
?
Jul 12, 2019 00:37
|
|
|
Cygni posted:fyi, the CPUs they used to send out were comedy option dead stock Bristol Ridge A6-9500's, and AMD was destroying them when you shipped them back. They make you ship it back so you have to actually go get a Zen CPU and don't just keep the terrible A6 and use that as the system. Giving someone a computer with an A6-9500 is a crime against humanity, so its probably for the best they are getting junked. Sounds like maybe they couldn't destroy them outright for whatever reason so someone came up with an idea to just mail them out to people for BIOS flashes and then they would be allowed to get rid of them.
|
|
#
?
Jul 12, 2019 00:51
|
|
|
3800x is in stock @ newegg for those that care. I'm in for one. EDIT: aaaaand they're gone Seamonster fucked around with this message at 05:18 on Jul 12, 2019 |
|
#
?
Jul 12, 2019 03:12
|
|
|
sativa dreams posted:I'm running this with a 3600. Zero issues so far. Hitting 4.2ghz and 3200CL14 with no problems. I'm surprised more people aren't all over this because: The Gigabyte boards are supposedly getting a BIOS update very soon that will let users change the chipset fan curves including even turning it off entirely.
|
|
#
?
Jul 12, 2019 04:29
|
|
|
xgalaxy posted:The Gigabyte boards are supposedly getting a BIOS update very soon that will let users change the chipset fan curves including even turning it off entirely. I don't know if it allows custom curves but they have a beta BIOS out with fan stop at least. https://old.reddit.com/r/Amd/comments/cby8og/aorus_x570_fan_stop_profiles/
|
|
#
?
Jul 12, 2019 04:49
|
|
|
Amazon finally delivered the Prime X570-P I ordered launch day, to go with the 3600X I've had since Tuesday. Haven't had much time to play with it, but here's a quick trip report: I missed entering UEFI setup on the initial boot to have it boot from my Windows 10 install USB so it booted to my existing install and...it works? It's a Windows 10 Pro install from a 6700k/z170 and everything seems to work just fine. I went ahead and installed the AMD chipset drivers and have yet to run into any obvious problems. I'll eventually nuke and pave it anyway because I know if I don't I'll eventually run into some mysterious problem that's impossible to troubleshoot, but for now I'm just going to leave it. The BIOS it comes with freezes on the confirmation dialog if you make any changes at all, gg Asus. Fortunately the launch day downloadable BIOS fixes it. It might have the "won't boost past 4.3GHz" issue. HWinfo says the current BIOS has AGESA 1.0.0.3 but I don't know if that's 1.0.0.3A or 1.0.0.3AB that's supposed to have the fixes and it's just truncating the letters, or if 1.0.0.3 is it's own version. With PBO enabled it sits at 4.2 exactly on all core loads, (which is normal even with the boost fixes, I think) and bounces around 4.3 on single core. I don't have the high idle voltage problem. Haven't had time yet to play around with the various PBO scaling and AutoOC options or try different voltage offsets. Currently running Prime95 Large FFT (don't want it to melt from small/blend while I'm not watching it) overnight. Under an NH-D15 it hasn't gotten above 75C yet. Edit: Forgot to mention that I don't really notice the chipset fan. It's audible, but just blends in with the other fans and background noise unless I'm actually trying to listen for it. Theris fucked around with this message at 05:10 on Jul 12, 2019 |
|
#
?
Jul 12, 2019 04:53
|
|
|
Day 5 of no 3900x from amd.com yet. Order at 10:15am on 7/7, order processed immediately, and had my CC charged. 5 days now without any shipping info. It honestly feels like they are sending their stock to retailers over filling online orders. I can't cancel the charge, the only thing to do is go through my credit card company and let them battle it out with AMD. I wouldn't be so bent out of shape if the website said back order. But it didn't. It said in stock. I've ordered ran from this build that was on back order and it finally shipped. I have no problems with stuff being on back order if I know what I'm getting into. This is a very lovely start from AMD and I'm not even using their product yet.
|
|
#
?
Jul 12, 2019 14:40
|
|
|
Zotix posted:Day 5 of no 3900x from amd.com yet. Order at 10:15am on 7/7, order processed immediately, and had my CC charged. 5 days now without any shipping info. It honestly feels like they are sending their stock to retailers over filling online orders. I can't cancel the charge, the only thing to do is go through my credit card company and let them battle it out with AMD. That sounds about par for buying direct from manufacturer. Whenever I buy anything from a manufacture I generally expect it to take 2 weeks to get to me but excellent support. I didn't even know AMD sold direct. If it's new there's a very good chance they are still not properly setup for direct sales fully. Business to Business it's perfectly fine to sell product that is on a container or not even in production yet. Then when stuff is delayed by a day or a week due to customs you let the other dozen parties know. When you sell direct people expect it to be sitting on a shelf ready to ship, which is almost never the case when buying directly.
|
|
#
?
Jul 12, 2019 15:18
|
|
|
|
| # ? May 31, 2024 22:31 |
|
|
Theris posted:I missed entering UEFI setup on the initial boot to have it boot from my Windows 10 install USB so it booted to my existing install and...it works? Since UEFI has become a thing the Windows bootloader is loving lightyears beyond where it was even in the Win7 days. Win10 is pretty goddamn resilient. I would advise you to actually do a clean install at some point, but you can absolutely do some test driving. I've flip flopped installs between Xeon e3s and Cascade Lake-SPs with absolutely no problem.
|
|
#
?
Jul 12, 2019 15:24
|
|
