|
If you don't do your delidding while wearing a heat suit inside an oven that I don't know what to tell you
|
#
?
Feb 20, 2017 19:55
|
|
|
|
| # ? May 15, 2024 02:00 |
|
|
eames posted:Asus motherboard leak seems to confirm DDR4 ECC support Fun, if you have ancillary stuff.
|
|
#
?
Feb 20, 2017 20:15
|
|
|
I went to some dying shopping mall today and while I was in their restrooms, I noticed someone used a knife to etch "AMD IS 2 FAST" into the stall door. Not sure what to make of it.
|
|
#
?
Feb 20, 2017 20:47
|
|
|
This guerilla marketing poo poo is getting out of hand.
|
|
#
?
Feb 20, 2017 20:50
|
|
|
buglord posted:I went to some dying shopping mall today and while I was in their restrooms, I noticed someone used a knife to etch "AMD IS 2 FAST" into the stall door. Not sure what to make of it.
|
|
#
?
Feb 20, 2017 20:54
|
|
|
Combat Pretzel posted:1x PCIe 16x, 1x PCIe 16x slot at 4x, 3x PCIe 1x 1 x PCIEx16, 1 x NVME (4x), USB etc - I could get by absolutely fine with that for ITX. Is this socket meant to last for 4 years worth of procs? I'm surprised they didn't add more lanes in that case. (I know sharings a thing, but still). GRINDCORE MEGGIDO fucked around with this message at 21:05 on Feb 20, 2017 |
|
#
?
Feb 20, 2017 21:02
|
|
|
I imagine we might see some updated motherboards in the meantime as well as CPUs. Just the socket will probably stay the same.
|
|
#
?
Feb 20, 2017 21:13
|
|
|
Combat Pretzel posted:1x PCIe 16x, 1x PCIe 16x slot at 4x, 3x PCIe 1x I think you missed one 16x slot. The only thing that would annoy me is the slow 1x M.2 slot (~985MB/s) 2x PCIe 16x slot (8x when both in use) 1x PCIe 16 slot at 4x 3x PCIe 1x slot 1x M.2 at 1x 8x SATA III USB 3.1/3.0/2.0: 2x/10x/6x more info at https://www.reddit.com/r/Amd/comments/5v4hqt/overview_ryzen_cpu_am4_mainboard_lineup_updated/ (picture of the board: https://cdn.videocardz.com/1/2017/02/ASUS-X370-PRIME-PRO-1.jpg)
|
|
#
?
Feb 20, 2017 21:19
|
|
|
How much overlap is there between the SATA/USB ports and the PCIe slots? On my current board, using the NVMe slot for instance disables the PCIe 4x slot. One PCIe 1x on the southbridge also goes, if I use certain SATA ports.
|
|
#
?
Feb 20, 2017 21:26
|
|
|
RyuHimora posted:???????? My Pentium G3258 right next to me was very definitely not soldered when I delidded it, in fact that's the reason I had to delid it, the thermal performance was alarmingly bad even on a really good cooler.
|
|
#
?
Feb 20, 2017 21:56
|
|
Anime Schoolgirl posted:weird, some people on reddit were reporting that it was soldered sometime in early 2015. i'm not sure why it was for those people, assuming they're not just making GBS threads us It sounds like BS to me, why would Intel make CPUs with a known cause of failure built in? I would assume the G3258 is a very small die CPU so it would definitely be affected by the problems with thermal cycling on small die soldered CPUs.
|
|
|
#
?
Feb 20, 2017 22:06
|
|
|
AVeryLargeRadish posted:It sounds like BS to me, why would Intel make CPUs with a known cause of failure built in? I would assume the G3258 is a very small die CPU so it would definitely be affected by the problems with thermal cycling on small die soldered CPUs. Taking a page from NVIDIA?
|
|
#
?
Feb 20, 2017 22:14
|
|
|
Bob Morales posted:Taking a page from NVIDIA? 
|
|
#
?
Feb 20, 2017 22:33
|
|
|
They wouldn't be any smaller than other mainstream i3-7 processors of that generation. Correct me if I'm wrong, but they all are just versions of the mobile chips (rather than Xeons which are the source of the HEDT or whatever chips) that either don't make the cut power wise, have other non-parametric defects like, say, a mask defect that results in a void where a couple transistors should be in one of the cores, or have stuff disabled intentionally since they sell more of the cheaper processors, and the yield is high enough that if they kept every die as the model that it is good for, there wouldn't be enough supply of the pentiums and i3/5s as well as an overabundance of i5/7s sitting in a warehouse somewhere. Now, iirc, those chips are small enough to where it wouldn't be advisable to solder them, but I could be wrong.
|
|
|
#
?
Feb 20, 2017 22:40
|
|
|
Watermelon Daiquiri posted:Why not just heat it up enough to melt the solder and weaken the adhesive? Too high of a temp? Or maybe you could carefully try to cut the adhesive and then run a heat gun over the cover to desolder the thing. Does intel/AMD sputter a sufficiently thick metal layer on the back of the die after they mill it down? I remember someone posting a video of a guy trying to do this with a blowtorch and an AMD CPU. The way I remember it going was the guy just cranking away at the heat spreader until he turned it around and noticed the pins had melted. edit: Might have been this one https://www.youtube.com/watch?v=P4Hp0xQhJwg Icept fucked around with this message at 22:45 on Feb 20, 2017 |
|
#
?
Feb 20, 2017 22:41
|
|
|
Combat Pretzel posted:How much overlap is there between the SATA/USB ports and the PCIe slots? On my current board, using the NVMe slot for instance disables the PCIe 4x slot. One PCIe 1x on the southbridge also goes, if I use certain SATA ports. Seems like the NVMe eats up some SATA and PCI-E ports from the CPU.
|
|
#
?
Feb 20, 2017 22:49
|
|
|
NVMe is literally a mechanism to directly attach storage devices to the PCIe bus, so yes, it'll eat PCIe lanes.
|
|
#
?
Feb 20, 2017 23:06
|
|
Icept posted:I remember someone posting a video of a guy trying to do this with a blowtorch and an AMD CPU. The way I remember it going was the guy just cranking away at the heat spreader until he turned it around and noticed the pins had melted. Heh, yeah a blowtorch of all things probably wouldn't be the best item to use though I'm not sure if it could reach a high enough temperature to melt the aluminum or whatever the actual pins are made from... Thankfully intel cpus are immune from that anyways
|
|
|
#
?
Feb 20, 2017 23:39
|
|
|
Watermelon Daiquiri posted:Thankfully intel cpus are immune from that anyways They're not immune from the poo poo stock TIM/IHS mount though. Really if AMD has their IHS soldered by default there isn't a reason to remove it (unless you're doing LN2 or phase change suicide OC's) since a soldered IHS will give you great temps. edit: you won't be melting pins on a AMD chip because of the soldered IHS = no reason to mess with IHS\/\/\/\/ PC LOAD LETTER fucked around with this message at 01:42 on Feb 21, 2017 |
|
#
?
Feb 21, 2017 01:33
|
|
PC LOAD LETTER posted:They're not immune from the poo poo stock TIM/IHS mount though. Really if AMD has their IHS soldered by default there isn't a reason to remove it (unless you're doing LN2 or phase change suicide OC's) since a soldered IHS will give you great temps. I was talking about immunity to melting pins
|
|
|
#
?
Feb 21, 2017 01:38
|
|
|
Let me tell you about pcie switches ,
|
|
#
?
Feb 21, 2017 01:59
|
|
|
Do we have any idea how Ryzen fares with encoding for watching 4K and streaming/recording?
|
|
#
?
Feb 21, 2017 10:10
|
|
|
ufarn posted:Do we have any idea how Ryzen fares with encoding for watching 4K and streaming/recording? nope.
|
|
#
?
Feb 21, 2017 10:25
|
|
|
https://www.youtube.com/watch?v=CcwJ6l33Zwo
|
|
#
?
Feb 21, 2017 12:13
|
|
|
This is a bit old cause I didn't follow this thread as closely as I should in the last three weeks but...mayodreams posted:AMD won't be getting into data centers with any significance for at least a couple of years. Brand new architectures and platforms need time to drivers sorted before trusting production level workloads. I have a raging AMD fan boy boner right now too, but typical hardware cycles are 3-5 years, and there will be a lot of hesitance on this new platform until real world data is available. The HPC people I'm talking to are standing in a loving line to get AMD into their datacenters right now, dumping intel and their CUDA apps, just to get on the Zen APU train a bit earlier, because big clients like CERN etc are breathing down their neck about it. What the poo poo are you on about? Paul MaudDib posted:I'm sure it'll be an overclocker's dream... A nightmare is a dream after all  Also, to all the people saying more cores won't see anything in gaming, I'm very glad I have a HT 4 core right now, it means I don't have to close my browser and other poo poo when playing games, doubly so in VR (as frame drops are even more annoying there), and now that games are slowly starting to move to 4 threads proper it's the perfect time to upgrade to a 6 or even 8 core, and keep that ability intact.
|
|
#
?
Feb 21, 2017 14:15
|
|
|
Josh Lyman posted:I, for one, want wood screws on my video card. No, son, what you need instead of wood screws are wooden screws. My videophile friends and I agree the wood gives the output picture a warmer feel and really embiggen the videostage produced by your monitor.
|
|
#
?
Feb 21, 2017 14:25
|
|
|
A Chinese website appears to have posted their R7 1700X review early. Some score comparisons are in the Latin alphabet for quick reference. Seems to beat out the i7 5960X in the benches they posted: http://archive.is/wh5HU Graphs:   Video Cardz also talk about it:https://videocardz.com/66182/amd-radeon-7-1700x-pictured-and-tested Choo choo 
|
|
#
?
Feb 21, 2017 14:38
|
|
|
Truga posted:
Why are they looking to dump their CUDA apps?
|
|
#
?
Feb 21, 2017 14:47
|
|
|
Because having GPU on package is so much better than having to run through a lengthy pcie bus. There's that CUDA -> OpenCL compiler now so they probably won't have to unless they run into any incompatibilities, I'm just saying that they would do it if they had to, to get to use Zen APUs.
|
|
#
?
Feb 21, 2017 14:49
|
|
|
ahh that makes a lot of sense actually.
|
|
#
?
Feb 21, 2017 14:59
|
|
|
The R7 1700X is basically identical (CB) or 5% faster (3DMark) compared to the 6900K at $389 vs $1050. I suspect that the overclocking headroom will be minimal because XFR is essentially automated overclocking but the value looks tremendous. I hope this is not a flash in the pan and AMD can keep this up without Keller on board.
|
|
#
?
Feb 21, 2017 15:49
|
|
|
NewFatMike posted:Graphs: Choo choo, Gonna be K7/K8 all over again.
|
|
#
?
Feb 21, 2017 15:55
|
|
|
If Intel doesn't drop prices across the board its going to be real interesting.
|
|
#
?
Feb 21, 2017 16:01
|
|
|
Sure hope Intel doesn't go OEM ratfucking again.
|
|
#
?
Feb 21, 2017 16:02
|
|
|
Wistful of Dollars posted:Sure hope Intel doesn't go OEM ratfucking again. Reminder: we are now under a Trump administration, with complete Republican control of government. Of course Intel will go back to their lovely ways with impunity.
|
|
#
?
Feb 21, 2017 16:08
|
|
|
SourKraut posted:Reminder: we are now under a Trump administration, with complete Republican control of government. Of course Intel will go back to their lovely ways with impunity. The EU has started also seriously enforcing antitrust laws in the last decade or two, so the imminent gutting of antitrust enforcement in the US won't get them off the hook completely.
|
|
#
?
Feb 21, 2017 16:19
|
|
|
evilweasel posted:The EU has started also seriously enforcing antitrust laws in the last decade or two, so the imminent gutting of antitrust enforcement in the US won't get them off the hook completely.
|
|
#
?
Feb 21, 2017 16:22
|
|
|
a pretty big german retailer lists a R1700/RX460 system for 999€ https://www.csl-computer.com/shop/product_info.php?products_id=13251
|
|
#
?
Feb 21, 2017 16:29
|
|
|
eames posted:a pretty big german retailer lists a R1700/RX460 system for 999€ That CPU/GPU pairing reminds me of the bad old days when PCs were sold with massively overpowered CPUs (for the time) but GPUs that were unbelievably weedy. All because most people only knew MORE MEGAHERTZ = BETTER!
|
|
#
?
Feb 21, 2017 16:38
|
|
|
|
| # ? May 15, 2024 02:00 |
|
|
Some more good numbers but it's still synthetics and unknown clock frequencies. http://wccftech.com/amd-ryzen-7-1700x-processor-tested/
|
|
#
?
Feb 21, 2017 16:40
|
|
