|
I came across a Gigabyte B450M motherboard that was not stable with the XMP profile provided by 2x 8GB 3600mhz sticks of Hynix C die ram. It turns out that the bios was ignoring the "I need 1.35v" part of the XMP profile, and decided to just go ahead and supply the ram with good ol' 1.20v JEDEC stock spec voltage instead. Despite running at advertised speeds / timings while being voltage starved, I could boot into Windows and not catch a blue screen until loading something heavy like Chrome. It was trivial to fix, but I had to go through some *serious* menu garbage to find the the setting I needed to change. The flow was something like "Power Management >> advanced settings >> power user settings >> advanced power user settings >> power user advanced power settings >> WARNING YOU ARE ABOUT TO VOID YOUR WARRANTY AND CAUSE FIRES >> I AGREE >> dram voltage >> (switched it from AUTO to MANUAL) >> YOU ARE WAY OUT OF LINE CHANGING THIS SETTING. TEAR WARRANTY IN HALF BEFORE PROCEEDING >> OK >> ARE YOU SURE YOU WANT TO DO THIS? >> YES >> (changed 1.20v to 1.35v) >> SAVE AND EXIT AND PROBABLY SCREW EVERYTHING UP >> OK. I rebooted, and it's been wonderfully stable at the correct timings and speed. But I can see how a lot of users would have thrown in the towel and RMA'd one or more things, further growing the legend of "RYZEN SO PICKY WITH RAM."
|
#
¿
May 15, 2020 23:43
|
|
|
|
| # ¿ Apr 29, 2024 12:29 |
|
|
CPU serial number being extended as a unique identifier via javascript to any website you visit would have been way worse than anything we have to deal with today. This was a "feature" for e-com that Intel and Microsoft were going to roll out, so if it had gained traction, it wouldn't be something you could gently caress with like you can with everything else used for browser fingerprinting. Edit: also, why do you think so much effort is put into tracking technology? If sites could just call navigator.GetIntelFreedomID then there would be no need for any abstract fingerprinting methods to exist. GRECOROMANGRABASS fucked around with this message at 23:30 on May 19, 2020 |
|
#
¿
May 19, 2020 23:23
|
|
|
Was curious if any B550 boards were in stock on Amazon this morning and wound up impulse buying a ASUS TUF Gaming B550-PLUS that will be here tomorrow. Will follow up with any positive or negative experiences.
|
|
#
¿
Jun 16, 2020 15:31
|
|
|
I noticed my motherboard touts 5,000 hour "long life" solid capacitors. Does anyone know what's up with that? 5k hours is ~208 days or around 7 months of 24/7 usage. I've never had electrolytic capacitors crap out on me that quickly, and I'd expect metal film capacitors to have a waaaay longer service life than that.
|
|
#
¿
Jun 18, 2020 19:07
|
|
|
ahh, that makes way more sense. Thanks guys!
|
|
#
¿
Jun 18, 2020 20:34
|
|
|
Got my B550 board in and setup. I was a little worried about my huge long graphics card over hanging the (fanless) southbridge heatsink, but temps don't seem to be an issue. Speaking of the southbridge, I noticed this in the ASUS licensed copy of CPU-Z:  I'm surprised to see "X570" as the identifier under SB. I've not been able to geek out on AMD news as much as I normally do lately, but I didn't think the X570 series and the B550 would share a chipset for some reason. Additionally, I'm confused by this chart on AMD's site, which seems to indicate that B550 doesn't support 2nd Gen (3000 series?) Ryzen processors such as the 3600 that I'm currently running in the B550 as I type this.  Source: https://www.amd.com/en/chipsets/x570
|
|
#
¿
Jun 21, 2020 02:38
|
|
|
Mr. Crow is correct. The youtube personas are putting out some ultra dumb hot takes to get your clicks. Just think for yourself.
|
|
#
¿
Jul 9, 2020 18:39
|
|
|
 That clearance height limit. 
|
|
#
¿
Aug 12, 2020 15:50
|
|
|
movax posted:I regret not buying into more TSMC when it was down in the $30s. Since this is the AMD thread, I'll point out that in the same time period you would have made more money from buying shares of AMD.. but yeah, I hold TSM too, and I'm kinda waiting for the market to react to the recent announcements. Surely the world's leading foundry is worth more than the ~$80 share price it's trading at. I guess the news about China hiring away some TSMC engineers has affected confidence; TSMC is known for having extremely loyal engineers with long tenures at the company. I really can't wrap my head around the motivation for the TSMC guys that left - on the upside, yeah, you are probably getting paid a lot more money than you were at TSMC. Downside is you now work for the CCP and they are investing an unfathomable amount of money with national security implications based on the expectation that you and a handful of other random engineers will be able to huddle up and spitball together a state of the art semiconductor foundry. edit: imagine being a poached TSMC employee showing up on your first day and finding yourself in a room full of washed out former Intel engineers arguing about the best path to 10nm. GRECOROMANGRABASS fucked around with this message at 16:16 on Aug 27, 2020 |
|
#
¿
Aug 27, 2020 16:03
|
|
|
IBM has been working on routing water through stacked chips as a cooling solution for ~12 years now. https://phys.org/news/2008-06-ibm-cools-d-chips.html More recent details: https://www.zurich.ibm.com/st/electronicpackaging/cooling.html It would be very interesting if they could pull it off and have a tiny processor resembling something from Sim Tower.
|
|
#
¿
Aug 31, 2020 13:03
|
|
|
sean10mm posted:I feel like we have zero idea what Zen 3 is going to do for performance per dollar, like not even to the level of the rumors leading to the RTX 3000 launch, which at least were mostly in the ballpark. Literally every tech blogger with "leaks" and "inside sources" about Zen 2 was proven to be full of poo poo. We didn't know anything about Zen 2 until Mama Su fed it to us straight from her beak, and it looks like Zen 3 will be the same.
|
|
#
¿
Sep 8, 2020 22:11
|
|
|
Blorange posted:The 5600x will be $300 because they can charge that much for it. It's looking like ~30% more performance for 50% more money. SiSoftware found performance to be ~40% better than the 3600X when reviewing the 5600X. And you are dead wrong about the price being "because they can charge that much for it." AMD is following the same pricing strategy for the 5000 series as Zen 2. Don't assume AMD has the same business model as Intel. Intel maintains a gross margin of around 65% on their consumer CPUs, while AMD forecasts expected gross margins on Zen 3 consumer CPUs to be roughly ~45%, which is where they have been for a while with Zen. Meaning AMD intends to price Zen 3 at the same level of per-unit profit as they received with Zen 2. Zen 3 has a lot of R&D to recoup, and is likely to be a more expensive product to produce due to factors such as TSMC capacity being at record high demand. They could charge a lot more than $299 for this processor. AMD buying Xilinx in an all-stock deal is a pretty clever move. Prior to the acquisition, AMD had a free cash flow of ~$150 million over the last year, but the combined companies will have nearly 2 billion in free cash flow over the next 12 months.
|
|
#
¿
Nov 2, 2020 22:36
|
|
|
Yeah, the alliance between TSMC, Samsung, and GF was great, and who knows where we might be with that combined R&D collaboration if GF didn't drop the ball. Samsung's 8nm process has a transistor gate pitch equal to or very close to TSMC's 10nm process, so I get the impression they aren't much further ahead of where they were when GF dropped out.
|
|
#
¿
Nov 2, 2020 23:29
|
|
|
Late stage capitalism is probably the reason for GF doing what it did. They had short and mid term profits guaranteed from their 12 and 14nm process nodes, and by having a nearly risk production ready 7nm process that just needs a little love and attention to get up and running, they probably have the exit strategy of being acquired for a king's ransom long before 12nm is obsolete.
|
|
#
¿
Nov 2, 2020 23:42
|
|
|
I ordered a Wraith Prism from Amazon to replace the dainty Wraith Stealth cooler that came with my 3600. I've never seen factory-applied thermal compound applied like this before. Is this a thing now, or should I be suspect of the authenticity of the merchandise? I intend to remove the stuff and replace with thermalright TF8 that I happen to have a tube of.. product claims heat transfer ability of 13.8 W/m.k. Any thoughts, comments, or goatse would be greatly appreciated. 
|
|
#
¿
Jan 16, 2021 01:29
|
|
|
Xaris posted:That seems too uniformly stamped on to be not-authentic. But I'd question the utility of them even doing that because it seems like it'd get dry or crappy. I've never heard of applying paste to a heatsink directly -- always cpu. I had the heatsink of an FX 8350 come with a solid square of paste, like .5mm thick before this, but yeah, the whole "thermal paste applied by dot matrix printer" aspect is weird to me. I appreciate your response, and I'll go ahead and answer your question, despite how vulnerable to being roasted the response will make me. I really like AMD, and when I look inside the tempered glass side of my case, I want to see two things: 1) An unsettling amount of over the top RGB 2) An AMD logo stamped onto a big rear end Karen of a heatsink looking back at me, approvingly GRECOROMANGRABASS fucked around with this message at 02:30 on Jan 16, 2021 |
|
#
¿
Jan 16, 2021 02:22
|
|
|
Thank you to all that responded. I removed the factory dot matrix paste from the heatsink, shined it up with some iso alcohol and coffee filters, and did the same with my 3600. Unfortunately, I discovered that the thermalright tf8 tube of thermal paste I had left over from a previous build was near empty - turns out that the syringe push depth is not a good indicator of how much paste is left. So I reached into the computer nostalgia box and found a little bit of Arctic Silver 5 from around 2008, a little bit more of Antec Silver from around the same period, and between the three tubes, managed to get a thin and even coating on the cpu, and an even thinner coating on the heatsink (I basically used a really small chunk to tint the surface of the cold plate) I'm averaging a solid ~10c below where I was with the stock cooler. Since booting, my min temp has been 26c, typical temp 33c, and max during gaming was briefly 61c, but a more typical gaming temp has been around 55c. I was going to buy more thermal paste, but gently caress it.
|
|
#
¿
Jan 17, 2021 19:36
|
|
|
BurritoJustice posted:I've got a 415v/32A three phase outlet in my basement as this house used to have an elevator and I'm imagining a three phase PSU to run a 7way setup I don't want to derail thread, but could you tell us more about this? Like, how many stories is your house? What is in the space where the elevator used to be? Did you find this by accident? I don't know why, but I fuckin' LOVE stories of homeowners finding poo poo they didn't know their house had behind like, a layer of drywall.
|
|
#
¿
Mar 4, 2021 16:20
|
|
|
Noobles posted:Well, out of the blue today my Asrock B550 just developed the USB issue after running fine for a month and a half. Hopefully they can figure something out with it soon. I’ve tried all the usual fixes and it’s still persisting. I had the USB device drop out issue ONLY when I had two mice plugged in and active at the same time. When I unplugged the ancient USB 1.0 mouse, the problem went away entirely. I also had pretty much all the USB ports populated, and USB protocols 1 through 3 gen 1 present. Further detail: I had a Logitech Unifying wireless receiver that a wireless keyboard & mouse connected to. In addition, I also had a USB wired keyboard and mouse connected. All USB devices would randomly crap out for about 15 seconds or so, and the only way to get an immediate fix was to unplug the old wired mouse, which immediately caused all other USB devices to function again. Before finding the solution of pulling the ancient Dell mouse or hearing of others having a USB drop out issue, I figured that the combination of Logitech unifying wireless receiver, Wifi 802.11 AC adapter, and a Bluetooth 5 nub all being within close proximity to one another was causing some kind of interference that was locking a driver, but I've had zero issues since getting rid of the ancient wired mouse. Motherboard is an Asus TUF Gaming B550-PLUS, BIOS version is 1401
|
|
#
¿
Mar 10, 2021 03:30
|
|
|
https://twitter.com/_markel___/status/1373059797155778562 https://twitter.com/_markel___/status/1373134141580058625 "No, it doesn't require physical connections... " implied remote microcode update? Haven't had my morning coffee yet, but drat INTEL, YOU NASTY. Did I fever dream it, or didn't intel drop like ~20 CVE at the end of February already? If a cloud provider like, I dunno, AWS were to see a major security breach through any number of Intel security calamities, I wonder how much legal exposure they would have at this point. How could a vendor reasonably argue that shared tenancy on Intel processors is a secure situation for their customers after endless bullshit like this being found?
|
|
#
¿
Mar 20, 2021 15:17
|
|
|
Yeah, the AMD drivers are great for multi-monitor/multi-resolution/multi-refresh-rate setups. I'm currently using a stock RX 480 blower model with 8GB of ram, and I use a 32" 4k 60hz monitor as a primary display with a secondary 27" 1440p 144hz monitor. Both monitors have an instance of Chrome with countless tabs open at the moment, and one of the monitors is playing a Hulu video. The power draw ranges from 34W with occasional, but infrequent bumps to ~46W. Edit: and one monitor is 10 bit color depth, the other 8 bit. One monitor shows FreeSync enabled, the other says "FreeSync Premium" is enabled. I'm pretty impressed with how smoothly this old video card manages two vastly different monitors without a hitch. GRECOROMANGRABASS fucked around with this message at 00:17 on Jul 2, 2021 |
|
#
¿
Jul 2, 2021 00:04
|
|
|
|
| # ¿ Apr 29, 2024 12:29 |
|
|
Klyith posted:[...]but if the economic effect is so big, why hasn't Intel done the same thing?[...] Say bud, I think it would make more sense if you thought about how Intel diligently perfected the art of slowly releasing minor incremental updates to their Pentium 4/Core-based architecture for well over a decade. Intel has made some great technology, but ... well you see what I'm getting at here.
|
|
#
¿
Mar 20, 2024 14:39
|
|