|
So, I bought a 1600 and I'm newish to both AMD (well, not really, owned a K6-2 once) and to overclocking. As in, only OCed my Intel chip last year new. I see that OCing Ryzen disables turbo and leaves me running at peak clocks all day long. So is there any point in OCing a 1600 to 3.7 (I'm on stock air so I don't want to push it) versus just leaving things at stock and letting turbo kick things to 3.6? I have yet to run benchmarks but it seems like without a big honkin' cooler this really isn't worth it.
|
#
?
Mar 8, 2019 23:46
|
|
|
|
| # ? May 16, 2024 15:29 |
|
|
Craptacular! posted:So, I bought a 1600 and I'm newish to both AMD (well, not really, owned a K6-2 once) and to overclocking. As in, only OCed my Intel chip last year new. Try OC'ing to an odd number like 3.675. For some reason it always runs a full speed on my system when it's an even number but not on odd. It's weird.
|
|
#
?
Mar 8, 2019 23:55
|
|
|
mcbexx posted:Eh, I have waited so long, I wanna see what the 570x chipset brings other than PCIe 4.0, do one big upgrade and be done with it for the next 5+ years (got the 2500k in... 2012?). Stopgapping with a 2700x for 4-6 months doesn't make sense to me. Just don't delay anymore pls. Also, don't gently caress up the launch.
|
|
#
?
Mar 9, 2019 00:08
|
|
|
Khorne posted:Probably not going to happen. DDR5 is coming, and so are massive improvements after zen2. Zen3 is supposed to be another zen2 tier jump, and intel's new stuff will probably be pretty good once it finally arrives in 2020/2021. The next few years are going to be chaos and nothing like 2011-2018. Is Zen3 supposed to be the last one that uses AM4 in any capacity? May just keep this 1600 until then. Praise the firesale. Edit: I basically chose to trade in my 3770K/Z77/DDR3 for a firesale CPU, the best VRMs you can get on a midrange AMD board, and 3200 RAM at prices that I'd call "still stupid". The hope is that I can throw something powerful in here eventually and stretch this mobo/memory's legs later. Hope this gamble pays off, because I'm not made of money!
|
|
#
?
Mar 9, 2019 00:18
|
|
|
If you're near a Micro Center they currently have the Ryzen 1600 on fire sale for only $80 (including cooler), presumably to clear them out before the Ryzen 3000 series launches. That's probably the best CPU deal ever.  EDIT: Just for fun I threw together a system including the 1600, a Gigabyte B350 board, 16 GB of DDR4-3000, an RX 570, a 256 GB SSD, a 500 watt PSU, and a case and it came out to only $420 before tax. That would be a great entry-level gaming computer. Or take out the graphics card and throw in a bunch of hard drives and that would make an excellent home storage server for under $300 before tax. The cheapest Intel 6-core CPU (without HT/SMT) they offer is $170.
Mr.Radar fucked around with this message at 00:29 on Mar 9, 2019 |
|
#
?
Mar 9, 2019 00:22
|
|
|
Klyith posted:I don't think it's totally a coincidence that Intel is seeing more vulnerabilities when their architecture is more complex. These vulnerabilities are effecting Intel CPU's going back over a decade, approaching nearly 20yr now really, to the early 2000's. I could see Intel having some sort of serious architectural advantage for a given generation to generation comparison vs AMD's parts (ie. Bulldozer vs Sandybridge) but over that time frame? Naaah. Why is it so hard for some people to believe that Intel just screwed up? Its not like their CPU's haven't had errata ever before. Same goes for everyone else this effects, some of which are AMD CPU's.
|
|
#
?
Mar 9, 2019 00:37
|
|
|
ASUS and MSI may have screwed up and released BIOS updates that indicate Ryzen 3000 is on the horizon. The last time this happened, Pinnacle launched within two months, and this is also AMD's 50th anniversary. IMHO, I think that makes for a May 1st paper launch, mostly announcement plus the Anniversary Processor with full launch later in the month. Navi seems to more close coincide with a Computex announcement but the rumored July launch.
|
|
#
?
Mar 9, 2019 01:07
|
|
|
It doesn't help that the techniques and sophistication of the teams that are hunting for these bugs has gotten substantially better over the last 10 years. The entire spectre class of bugs and sidechannel leakage detection really only became a thing in academia a few years ago. The fact that Intel was the first one to have it detected probably had more to do with it's install base than anything else. Easy to do research on a uarch when it's sitting in your lovely laptop vs. needing some fpga dev board to poke it.
|
|
#
?
Mar 9, 2019 01:28
|
|
|
Khorne posted:Probably not going to happen. DDR5 is coming, and so are massive improvements after zen2. Zen3 is supposed to be another zen2 tier jump, and intel's new stuff will probably be pretty good once it finally arrives in 2020/2021. The next few years are going to be chaos and nothing like 2011-2018. So even if I upgrade from my 2500k to a 8C/16T Zen2 CPU something much better is going to be right around the corner (1-2 years) anyway? Now I want CPU tech to stagnate again so my next build will last as long at my 2500k build has lasted. I don't want to build a 2700X or a 3700X rig this Fall and then have to upgrade again only 2-3 years later. Edit: but if my 2500k rig dies before then, gently caress it, I'll build a cheap 2600X system and upgrade again in 3 years. spasticColon fucked around with this message at 02:26 on Mar 9, 2019 |
|
#
?
Mar 9, 2019 02:20
|
|
|
There is always 'something much better' 1-2 years away.
|
|
#
?
Mar 9, 2019 02:28
|
|
|
Mr.Radar posted:If you're near a Micro Center they currently have the Ryzen 1600 on fire sale for only $80 (including cooler), presumably to clear them out before the Ryzen 3000 series launches. That's probably the best CPU deal ever. I saw this and have been real tempted to throw together a gaming / Adobe CC system with either a b450 or x470 motherboard, but I'm not sure what the next gen chipset will bring. I've got an i5 4670k but I think I can just wait at this point. Still a hell of a deal. GuyonthecoucH fucked around with this message at 03:03 on Mar 9, 2019 |
|
#
?
Mar 9, 2019 02:39
|
|
|
GuyonthecoucH posted:but I'm not sure what the next gen chipset will bring. Otherwise the chipset is based on all the IO stuff (SATA, USB 3.x, extra PCIe lanes, etc) that is integrated into the die of Epyc and TR. Methylethylaldehyde posted:Easy to do research on a uarch when it's sitting in your lovely laptop vs. needing some fpga dev board to poke it.  But OEM's usually put AMD chips in the shittiest of laptops, particularly during the Bulldozer era, so  by your logic AMD chips should be extra super easy to do research on!  Yeah Intel's stuff is ubiquitous so you guys are teeeeeeeechnically right that its easier to do research on them but realistically the AMD chips are also pretty common and easy enough to get (arguably easier to get if cost is the biggest barrier you face) that no real barriers exist to doing work on them either.
|
|
#
?
Mar 9, 2019 03:08
|
|
|
taqueso posted:There is always 'something much better' 1-2 years away. I actually don't want stagnation but I don't want to have to upgrade my rig every 2-3 years either. My rig before this one my C2D E8400 lasted me only three years but my current rig my 2500k@4.2GHz has lasted me almost eight years now and still runs every game I throw at it but some newer games don't run as good as they could because the CPU is finally starting to show its age.
|
|
#
?
Mar 9, 2019 03:09
|
|
|
spasticColon posted:So even if I upgrade from my 2500k to a 8C/16T Zen2 CPU something much better is going to be right around the corner (1-2 years) anyway? More bandwidth really depends on your workload. Most games don't scale extremely well with bandwidth, and we have yet to see how Infinity Fabric scales with the i/o on a separate chip. Zen 2 could be a 9900k for half price or similar price and many more cores, as well as lower tdp for those that care about that. fake e: also non validated ecc support
|
|
#
?
Mar 9, 2019 04:08
|
|
|
I'm super happy if my pc is totally outclassed in 2 years. Bring it on. You don't "have" to upgrade anything.
|
|
#
?
Mar 9, 2019 04:11
|
|
|
spasticColon posted:So even if I upgrade from my 2500k to a 8C/16T Zen2 CPU something much better is going to be right around the corner (1-2 years) anyway? But I definitely wouldn't over spend with the idea that sandy/ivy bridge is "normal". Look at all the people who bought non-k CPUs and got caught in intel's vicious planned obsolescence cycle. Or anyone who bought the pretty bad AMD processors at the time. Most people make frequent, poorly planned impulse buys when it comes to CPU parts. There are people who literally bought a 1050, 1060, 1080 in the same generation. You will likely still be able to run everything for a long time on zen+/8700k/9 series intel CPUs, but the entire point of upgrading is you're getting notably more performance. It's why I stayed on the 3770k - the 8700k was the first real upgrade. I just see the market as having appealing upgrades over the next few years unlike from 2011 to the 8700k release. The only big consideration is zen is still relatively immature. They have lots of changes coming to the architecture. And Intel has been sleeping on a lot of its improvements because 10nm was supposed to be here like 5 years ago. On top of that, 5nm should be doable on EUV which should finally go into good-yield production next year with 7nm+ and intel's 10nm+. Khorne fucked around with this message at 04:50 on Mar 9, 2019 |
|
#
?
Mar 9, 2019 04:43
|
|
|
spasticColon posted:I actually don't want stagnation but I don't want to have to upgrade my rig every 2-3 years either. My rig before this one my C2D E8400 lasted me only three years but my current rig my 2500k@4.2GHz has lasted me almost eight years now and still runs every game I throw at it but some newer games don't run as good as they could because the CPU is finally starting to show its age. R5 2600/B450/16GB DDDR4 is only ~$300 with deals and its still gonna take forever to become semi-obsolete in games. If you can't even afford that you have more serious IRL issues than PC gaming.
|
|
#
?
Mar 9, 2019 04:44
|
|
|
Khorne posted:Probably not going to happen. DDR5 is coming, and so are massive improvements after zen2. Zen3 is supposed to be another zen2 tier jump, and intel's new stuff will probably be pretty good once it finally arrives in 2020/2021. The next few years are going to be chaos and nothing like 2011-2018. The first wave of DDR5 is likely to go like the last few gens, barely faster than the top of the line of last gen, just less voltage. No way affordable sticks are going to be a significant upgrade over 3200CL12/14 DDR4 until several years from now.
|
|
#
?
Mar 9, 2019 05:46
|
|
|
Palladium posted:R5 2600/B450/16GB DDDR4 is only ~$300 with deals and its still gonna take forever to become semi-obsolete in games. If you can't even afford that you have more serious IRL issues than PC gaming. I mentioned that I plan on doing that if my current rig dies before Zen 2 launches. quote:Edit: but if my 2500k rig dies before then, gently caress it, I'll build a cheap 2600X system and upgrade again in 3 years. spasticColon fucked around with this message at 07:10 on Mar 9, 2019 |
|
#
?
Mar 9, 2019 06:45
|
|
|
sauer kraut posted:The first wave of DDR5 is likely to go like the last few gens, barely faster than the top of the line of last gen, just less voltage. Give it 1 or 2 years at least after 1st release for DDR5 to mature, drop in price, and for faster enough to matter DIMMs to come out and then it'll be worth it. Here is what 2x 8GB DDR4 RAM prices and speeds were like near launch back in 2014 as a reminder of what to expect DDR5 launch prices to be like:  I think the only potential gotcha is if the DRAM foundries decide to do some price fixing again but that sort of thing can happen at any time.
|
|
#
?
Mar 9, 2019 07:21
|
|
|
For the same reasons, hopefully AMD is smart enough to design AM5 with hybrid DDR4/DDR5 support (preferably both on a same mobo) to steal more Intel's thunder.
|
|
#
?
Mar 9, 2019 07:41
|
|
|
GRINDCORE MEGGIDO posted:I'm super happy if my pc is totally outclassed in 2 years. Bring it on. You don't "have" to upgrade anything. I definitely agree with this. Stagnation just feels horrible overall. I know the drum has been beat many times on this but we've felt this stagnation before with 2500k to 6600k being barely much of a jump and of course nvidia's & AMD's total failure to advance a generation decently in the GPU arena. Ryzen is the best thing to happen in years
|
|
#
?
Mar 9, 2019 11:29
|
|
|
PC LOAD LETTER posted:
I don't know how it is in other universities, but in my university it would require quite a bit off extra effort. The process of purchasing a computer is arduous enough that we need 4 or 5 year warranties. That has limited us to HP or Lenovo high end business laptops and they've never offered one with AMD CPUs. If some research wants an AMD they either have to pay it from their own pocket, or they will have to go through the bureaucracy and justify why they need to buy something outside our supply contracts.
|
|
#
?
Mar 9, 2019 13:30
|
|
|
Saukkis posted:That has limited us to HP or Lenovo high end business laptops and they've never offered one with AMD CPUs. Business oriented Lenovo AMD laptops are out there too though they're not high end something like that would probably have a decent shot at making it through the bureaucracy. Really though if you're university is going to place weird restrictions (ie. "you can only buy ~$2k+ "workstation" laptops and no desktops ever") on what you can get that is whole other issue that doesn't have anything to do with availability in general of AMD parts.
|
|
#
?
Mar 9, 2019 14:52
|
|
|
PC LOAD LETTER posted:"High end" is a bit subjective in terms of cost and features but business HP AMD laptops with good specs and build quality do exist and just takes a few clicks on their site to get them. The models available to us don't have AMD CPUs. We have gotten them through public tender and we only get the models Lenovo decided to offer in the tender. I don't know our required specs in the recent tender, but I doubt they would have specified Intel CPUs, but they probably would have done something to exclude Bulldozers at least. I think we got a pretty good offer. 715€ for Thinkpad T480 with i5-8250U, 16GB of RAM, 512GB SSD, 14" 1920x1080 and 4 year warranty feels like a pretty great price for me. I checked a large online store and their closest match would cost 1492€ with 256GB SSD and 3 year warranty.
|
|
#
?
Mar 9, 2019 16:16
|
|
|
Again if they're gonna place weird restrictions on what you can buy that is a pretty separate issue altogether from what was talked about before in thread on this subject. Having limited options due to weird buying restrictions is also pretty different from "these things don't exist anywhere" too.
|
|
#
?
Mar 9, 2019 16:39
|
|
|
My adventures in overclocking on Ryzen continue. I finally nutted up and put the 1.46v that my memory apparently needs to run at DDR4-3600 speeds according to Ryzen DRAM Calculator and sure enough it's doing that. From what AMD engineers have said and what I've read, apparently up to 1.5v is fine for daily driver OC's on the memory and it doesn't seem like you can get much past DDR4-3200 on Ryzen without increasing voltage past the normal 1.35v XMP recommends in its timings. The increase in overall performance on this 2700X + DDR4 3600 14-15-15-30 compared to my old 1700X + DDR4 3000 CL16 is pretty dramatic in previously bottlenecked situations, seeing over a 20% increase in performance in CPU bottlenecked situations. Haven't done a proper full stability test yet, but it did survive doing Cinebench R15/R20 as well as various game benchmarks. Anything past DDR4-3600 also seems not to be worth doing, you really need to start pushing the latency up and going past the 1.5v AMD recommends. In WoW, there is a little ledge I can stand on that consistently gave me 40 FPS on my old setup looking out into the horizon, it now does 52 FPS in the same spot, that was exactly the type of increase in performance I was looking for to shore up those bad low end spikes.
|
|
#
?
Mar 9, 2019 16:40
|
|
|
PC LOAD LETTER posted:Again if they're gonna place weird restrictions on what you can buy that is a pretty separate issue altogether from what was talked about before in thread on this subject. No, I think this is exactly what was talked about. Let's guess that there is a 95% chance the laptop the researcher is using every day will have an Intel CPU inside. Also their previous laptop still laying on the nearby shelf probably is Intel. Same as all the other laptops and desktops the research group has laying around to tinker with. When the researcher wants to do some quick test it's easiest to use the machines nearest to them. After the researchers have found these vulnerabilities on Intel CPUs, they already have all the material for a great research paper. They can try to figure out if AMD or ARM has the same vulnerabilities, but that doesn't significantly improve their paper and it will delay the publication, so Intel-only is good enough. Some other research group could continue their work and test other microarchitectures, but their research paper would not have nearly the same impact the original paper had, so it would only be worth doing if they can't come up with anything more original. Speaking of procurement, I wouldn't call the restrictions weird. We are a public institution so we are bound by the EU procurement laws, and through us our suppliers are also bound by them. Public procurement is an annoying hassle, but thanks to them we do get great prices.
|
|
#
?
Mar 9, 2019 17:00
|
|
|
PC LOAD LETTER posted:Yup. Its almost never worth it to buy a new system just to switch to the latest RAM when its first released. Maybe one day I can get 16GB of DDR4 for less than I paid for 16GB of DDR3 7 years ago
|
|
#
?
Mar 9, 2019 17:06
|
|
|
Saukkis posted:No, I think this is exactly what was talked about. Let's guess that there is a 95% chance the laptop the researcher is using every day will have an Intel CPU inside. Also their previous laptop still laying on the nearby shelf probably is Intel. Same as all the other laptops and desktops the research group has laying around to tinker with. When the researcher wants to do some quick test it's easiest to use the machines nearest to them. Finding the vulnerability in AMD or ARM processors would greatly enhance the profile of the security researcher and their team. Even if you insist the publishers cannot get access to AMD processors (...) and therefore failed to find the same exploits in AMD chips, you can bet your marbles that AMD and ARM processors have been heavily audited for similar flaws by peers of the publishers; it would be low hanging fruit to find a similar attack vector in other processors once the details of the fundamental security flaws and POC for Intel processors was published. Lastly, there is capitalism. Anyone with money in Intel - including Intel itself, has profit motive to throw money at security researchers with the purpose of finding security vulnerabilities in AMD processors to mitigate PR outrage. If a team of security researchers CAN publish something that might help turn "Intel security flaws" into "modern CPU security flaws" they can expect a big pay day. In fact, there was already an attempt at this last year, where "exploits" were published with catchy names for AMD platforms that included crap like "if a third party gets physical possession of the motherboard.." and "if the user can be tricked into downloading a malicious bios image (that has somehow been signed by the motherboard manufacturer) and flashes the bios with it.. bad stuff COULD happen. Just seeing that was the best they could come up with was reassuring tbh.
|
|
#
?
Mar 9, 2019 19:45
|
|
|
Guys, I cannot take this seriously. Do you really think that the types of security researchers who do the heavy-hitting poo poo that causes a CPU to lose control of it's sphincters are sitting around with only their own work laptop to experiment on? CPUs are not that expensive! People at reputable institutions get stuff provided at big discounts! Like, if that were really the case why is Intel not shipping a pallet of AMD chips to all of these teams? They could spend a million dollars outfitting every academic security research team in the world with a selection of processors from every major architecture & family. If the result was that AMD and ARM were getting just as much bad news as Intel, that would be worth a billion dollars to them. Also, one of the two teams that independently discovered meltdown was Google Project Zero. Believe whatever you want about the edu guys, if you think Google can't afford a couple ryzen CPUs for their elite team you're a moron.
|
|
#
?
Mar 9, 2019 21:00
|
|
|
As a person who simply buys whatever is best for my price range and doesn't get caught up in business wars, I'll never understand the people who take it to the level of thinking that chipmakers are using security research firms to sabotage one another etc. It's not that it can't happen, it's just that it's such a waste of time. It's an incredibly targeted level of marketing that costs quite a bit of money to reach very few people, because very few people are reading security whitepapers. Even if AMD bought an Intel flaw, it doesn't matter because Intel will just make another prime time TV blitz with another Blue Man Group or Jim Parsons or whomever. Like your mom is not going to understand speculative prediction memory addressing like she's going to understand Kylie Jenner talking about her laptop. EDIT: it also helps that, thanks to the shift to battery-powered devices, most people don't choose AMD or Intel. They choose Dell or Lenovo or Apple. Enthusiast builders are where the short term money is, not the long term money. If you lose enthusiast builders, you can get them back by fixing your poo poo and winning either performance or price:perf. Craptacular! fucked around with this message at 21:34 on Mar 9, 2019 |
|
#
?
Mar 9, 2019 21:26
|
|
|
spasticColon posted:I actually don't want stagnation but I don't want to have to upgrade my rig every 2-3 years either. My rig before this one my C2D E8400 lasted me only three years but my current rig my 2500k@4.2GHz has lasted me almost eight years now and still runs every game I throw at it but some newer games don't run as good as they could because the CPU is finally starting to show its age. I'm still impressed by the number of people who are still pushing those Sandy Bridge-era rigs without much complaint. I could have stuck with my Ivy Ridge 3570k rig for another year or two, but I was really getting bored of it after nearly 6 years, plus I was worried about the mobo dying at some point*. *A couple of years ago, the old PC took a lightning strike that killed the on-board ethernet, the cable modem and the PSU. I'm still amazed the mobo didn't up and croak right then and there.
|
|
#
?
Mar 9, 2019 22:52
|
|
|
I ran a 965 BE for drat near eight years, and it's still doing work as a hand-me-down. If it makes it to 2020 without dying, I'm repossessing the computer from the stepbrothers, and giving them new parts, so that I can enshrine the 965 BE in a place of honor. As far as I'm concerned, Sandy Bridge is the Johnny-come-lately to Deneb's Methuselian longevity. (Of course, it didn't hurt that everything between then and Zen sucked rear end, but still.) SwissArmyDruid fucked around with this message at 23:23 on Mar 9, 2019 |
|
#
?
Mar 9, 2019 23:10
|
|
|
Any idea if / when an AMD apu will be able to playback 4k Netflix video?
|
|
#
?
Mar 9, 2019 23:50
|
|
|
GRINDCORE MEGGIDO posted:Any idea if / when an AMD apu will be able to playback 4k Netflix video? When the GPU side submits to PlayReady 3.0 I guess? If it's separable from the CPU side, anyway. I don't think AMD's in any rush to pull an SGX.
|
|
#
?
Mar 9, 2019 23:59
|
|
|
GRINDCORE MEGGIDO posted:Any idea if / when an AMD apu will be able to playback 4k Netflix video? does the netflix/intel deal allow for other cpus/gpus to play 4k?
|
|
#
?
Mar 10, 2019 00:05
|
|
|
wargames posted:does the netflix/intel deal allow for other cpus/gpus to play 4k? Nvidia's GPUs (from Pascal, anyway) reportedly do it, but they submit to PlayReady 3.0. e: Which is weird because Vega is supposed to support PlayReady 3.0, and Microsoft is supposed to have worked with AMD on it. It could be straight exclusion? e2: this AMD community thread suggests it got supported driver-side but only for discrete Polaris. And not forward-ported to Vega. What. dont be mean to me fucked around with this message at 00:23 on Mar 10, 2019 |
|
#
?
Mar 10, 2019 00:07
|
|
|
Craptacular! posted:As a person who simply buys whatever is best for my price range and doesn't get caught up in business wars, I'll never understand the people who take it to the level of thinking that chipmakers are using security research firms to sabotage one another etc. You're ignoring the business market. Someone responsible for purchasing thousands of CPUs may still not read white papers, but they'll probably have read the write ups that will be all over the tech sites.
|
|
#
?
Mar 10, 2019 00:29
|
|
|
|
| # ? May 16, 2024 15:29 |
|
|
I'm sure quite a few data centre operators have many people concerned about those papers. Efb
|
|
#
?
Mar 10, 2019 00:58
|
|