|
Its probably real its just not the chip people think it is. I seriously doubt thats going to be the cheap $100 buck one
|
#
?
May 26, 2019 15:23
|
|
|
|
| # ? Jun 7, 2024 00:16 |
|
|
Cao Ni Ma posted:Its probably real its just not the chip people think it is. I seriously doubt thats going to be the cheap $100 buck one 24600. Microcenter is selling them for $129.99 with $30 off motherboard. The 2600 is priced at $159.99 with $30 off motherboard. It scores around 5000 / 24000 (maybe less multicore, looking through geekbench results is somewhat difficult, multi varied from 17000-25000 with that single core score) with good RAM and a good OC as well. It's hard to say without knowing the details, but that's very much a sub $200 CPU. I wouldn't bet on $100 but who knows what AMD is up to. If there is a large clock discrepancy between that processor and the more expensive ones I wouldn't rule out a pretty low price point. With Intel not on a competitive node they can get away with charging more, though. Khorne fucked around with this message at 15:53 on May 26, 2019 |
|
#
?
May 26, 2019 15:40
|
|
|
I hold by my bet that this refresh won't bring us more cores per dollar, and the 12 and 16 core ones will be $400 and $500 parts while the 4-8 core parts hold the same price point.
|
|
#
?
May 26, 2019 15:40
|
|
|
Twerk from Home posted:I hold by my bet that this refresh won't bring us more cores per dollar, and the 12 and 16 core ones will be $400 and $500 parts while the 4-8 core parts hold the same price point. Still the fact we're having this discussion at all is pretty absurd value. If the clocks go near where they claim, there's literally no IPC improvements and prices line up like that it's still excellent value. If you're right and the 12 is $419 at the hype level it's still pretty fantastic. I just hope navi is good enough it moves the market a little. Wanting a 2080 to reach my performance targets is just an unpleasant amount of money.
|
|
#
?
May 26, 2019 15:52
|
|
|
Twerk from Home posted:I hold by my bet that this refresh won't bring us more cores per dollar, and the 12 and 16 core ones will be $400 and $500 parts while the 4-8 core parts hold the same price point. Even if prices are aimed that low, retailers will make sure the new hot poo poo everyone has been waiting for will be marked up significantly.
|
|
#
?
May 26, 2019 16:07
|
|
|
Gamers Nexus says Ryzen 3000 announcement is June 10th and that we should only expect X570 launch now, preorders will be on July 1st, July 7th will be the launch date. Supposedly the 16 core chips will hit 300W power consumption with air cooling. https://www.youtube.com/watch?v=PEAVSAoC_Tg
|
|
#
?
May 26, 2019 16:30
|
|
|
mcbexx posted:Even if prices are aimed that low, retailers will make sure the new hot poo poo everyone has been waiting for will be marked up significantly. That's what I'm worried about. Even if there is enough supply to meet demand at launch retailers are probably still going to price gouge the first few months after launch especially if Zen 2 truly lives up to the hype. Edit: Blueballed until June 10th. 
spasticColon fucked around with this message at 16:43 on May 26, 2019 |
|
#
?
May 26, 2019 16:34
|
|
|
Measly Twerp posted:Gamers Nexus says Ryzen 3000 announcement is June 10th and that we should only expect X570 launch now, preorders will be on July 1st, July 7th will be the launch date. Only officially supporting DDR4-3200 is pretty ugh, hopefully at least 3600 is reliable in practice.
|
|
#
?
May 26, 2019 16:43
|
|
|
K8.0 posted:Only officially supporting DDR4-3200 is pretty ugh, hopefully at least 3600 is reliable in practice. That's 100% normal, though. Current Zen CPUs support a maximum of DDR4-2933. And the i9-9900K stops at DDR4-2666. Processor manufacturers never support overclocked memory in an "official" capacity. You'll be able to run 3600 just fine. Lambert fucked around with this message at 16:55 on May 26, 2019 |
|
#
?
May 26, 2019 16:52
|
|
|
Measly Twerp posted:Gamers Nexus says Ryzen 3000 announcement is June 10th and that we should only expect X570 launch now, preorders will be on July 1st, July 7th will be the launch date. See, if I had a killer product sitting in my drawer that I know people are dying to get their hands on or at least have some official product details to dig into I would get that info out right the gently caress now. If not just to keep the hype going and stop potential customers from saying "gently caress it, I'm tired of waiting, I'm going Intel now". It's not like specs are going to change drastically at this point in the process, right? I am very slowly suspecting Ryzen 3000 will be another AMD wet fart. Too little, too late. Maybe too hot and power hungry. Or way pricier than we hope for. You heard it here first, folks. Brace for disappointment.
|
|
#
?
May 26, 2019 17:29
|
|
|
mcbexx posted:See, if I had a killer product sitting in my drawer that I know people are dying to get their hands on or at least have some official product details to dig into I would get that info out right the gently caress now. It could honestly be marginally better than Zen+ in every way and still be a great product, in my mind. Everyone has sky-high expectations, but that doesn't mean a product that performs only slightly better than Zen+ wouldn't be a welcome addition to the market.
|
|
#
?
May 26, 2019 17:56
|
|
|
Measly Twerp posted:Supposedly the 16 core chips will hit 300W power consumption with air cooling. If it's true, I wonder what that'll mean for the eventual 32C TR ones.
|
|
#
?
May 26, 2019 17:57
|
|
|
Combat Pretzel posted:Eh? A 16C Threadripper does way less than that. The 300 watts figure is from an overclocked part on a high-end AIO water cooler. It would surprise me if AMD released an AM4 part with a stock TDP over 125ish watts (just from the perspective of what kind of stock cooler they'd have to include).
|
|
#
?
May 26, 2019 18:08
|
|
|
Wasn't July 7th the announced date like, over a hundred days ago?
|
|
#
?
May 26, 2019 18:15
|
|
|
Spiderdrake posted:Wasn't July 7th the announced date like, over a hundred days ago?
|
|
#
?
May 26, 2019 18:19
|
|
|
AdoredTV backpeddling from his rumors on announcement day  https://twitter.com/AdoredTV/status/1132664835773259777
|
|
#
?
May 26, 2019 18:49
|
|
|
MaxxBot posted:AdoredTV backpeddling from his rumors on announcement day Covering every possible angle so you're always correct
|
|
#
?
May 26, 2019 18:59
|
|
|
Lambert posted:That's 100% normal, though. Current Zen CPUs support a maximum of DDR4-2933. And the i9-9900K stops at DDR4-2666. With the slower timings on faster memory, past about 2400 or 2666 it's often slower or equal performance at higher cost. Check some benchmarks.
|
|
#
?
May 26, 2019 19:06
|
|
|
Yeah, higher freq ram was nice when infinity fabric ran at ram speed, since zen2 is apparently free of that bullshit so just get the cheap ram.
|
|
#
?
May 26, 2019 19:24
|
|
|
Broken Machine posted:With the slower timings on faster memory, past about 2400 or 2666 it's often slower or equal performance at higher cost. Check some benchmarks. Absolutely not even close to true, and never has been. The goon advice to just buy slow ram has been bad advice going back to at least Sandy Bridge and you still see people giving it. There are actually a lot of games that benefit a lot from higher memory speed, and consistently more of them as time goes on. Especially with ram prices having come down a ton, it's foolish to be telling people to cheap out. 20 bucks for a 5-10% performance improvement is a really good deal. e - and to be clear, this is platform agnostic. If anything it applies more on Intel because their CPUs are faster and are able to chew through data faster. K8.0 fucked around with this message at 19:39 on May 26, 2019 |
|
#
?
May 26, 2019 19:34
|
|
|
K8.0 posted:Absolutely not even close to true, and never has been. The goon advice to just buy slow ram has been bad advice going back to at least Sandy Bridge and you still see people giving it. There are actually a lot of games that benefit a lot from higher memory speed, and consistently more of them as time goes on. Especially with ram prices having come down a ton, it's foolish to be telling people to cheap out. Ok then, prove it. Show me some real actual benchmarks showing significantly increased performance. Under certain cpu intensive uses, you may see slightly more performance. For most users under typical workloads, you will not notice an improvement that is worth the cost, especially as you go up towards the highest frequencies. With the increase in latency you may actually have less bandwidth. The benefits are marginal at a considerable cost premium. Here is one of myriad benchmarks showing it doesn't matter. Oh wow, a 4% multicore increase, and otherwise flat performance.
|
|
#
?
May 26, 2019 19:45
|
|
|
Broken Machine posted:With the slower timings on faster memory, past about 2400 or 2666 it's often slower or equal performance at higher cost. Check some benchmarks. That might be true for 3200, but it sure as hell isn't true for 2666. You just can't tighten the timings enough to match 3200@CL14.
|
|
#
?
May 26, 2019 19:46
|
|
|
Broken Machine posted:Ok then, prove it. Euer Wunsch ist mir Befehl: https://www.computerbase.de/2019-03/amd-ryzen-cpu-ddr4-ram/2/?amp=1 The 99th precentile is the important bit.
|
|
#
?
May 26, 2019 19:54
|
|
|
The difference between 2666 and 3000 can be $5. The difference between 3000 and 3200 is often like $10. It’s much less than the dumb colored lights people often buy. Corsair once either wrote or sponsored some kind of “speed vs timings” article that concluded more speed is better. You can tinfoil that they just want consumers to spend more, but the cost difference is the cost of an SA account or less. People post bad and re-reg all the time, are you gonna quibble about tenbux in a part you’ll use for years?
|
|
#
?
May 26, 2019 19:54
|
|
|
Measly Twerp posted:Gamers Nexus says Ryzen 3000 announcement is June 10th and that we should only expect X570 launch now, preorders will be on July 1st, July 7th will be the launch date. Steve said it is 300W is when overclocked using non exotic cooling, so with a big air cooler or liquid at ambient temps. Seems reasonable for an overclocked 16 core part. https://www.reddit.com/r/hardware/comments/bt93ji/gamers_nexus_ryzen_3000_release_date_oc_power/eov32pc/ B-Mac fucked around with this message at 19:57 on May 26, 2019 |
|
#
?
May 26, 2019 19:55
|
|
|
So 150W for a 8C at 4.5 GHz? Sounds... shockingly 14nm.
|
|
#
?
May 26, 2019 19:58
|
|
|
If the June 10th info is true then I don't know WTF the AMD marketing team is thinking. First you had expectations of a CES reveal, then you had 5 months of Computex talk. It's only a couple more weeks, until June 10th, but at this point it's kind of frustrating to people who have held off on purchases for several months.
|
|
#
?
May 26, 2019 20:02
|
|
|
Broken Machine posted:Ok then, prove it. Show me some real actual benchmarks showing significantly increased performance. https://www.techspot.com/article/1171-ddr4-4000-mhz-performance/page3.html For one. Note that this is on a 4.5ghz skylake, on a 5ghz 9700k/9900k you'd see even more scaling. Also I can't be hosed to dig them up, but several people have done benchmarks showing how if you bought 1866 mhz ram with your Sandy Bridge, 5 years later you had like a 10%+ average performance advantage over people who went with 1333, which was like half the gap to new hardware at the time. Games have been consistently becoming more and more bandwidth hungry for a long time, and while the bottlenecks rarely show up when a system is new, they become dramatic over time. There's a reason even the slow-to-update parts picking thread OP has recommended everyone get 3200+ for a while now.
|
|
#
?
May 26, 2019 20:02
|
|
|
BeastOfExmoor posted:If the June 10th info is true then I don't know WTF the AMD marketing team is thinking. First you had expectations of a CES reveal, then you had 5 months of Computex talk. It's only a couple more weeks, until June 10th, but at this point it's kind of frustrating to people who have held off on purchases for several months. I was serious earlier, if Lisa blueballs the crowd again she's going to get booed off the stage.
|
|
#
?
May 26, 2019 20:02
|
|
|
Craptacular! posted:The difference between 2666 and 3000 can be $5. The difference between 3000 and 3200 is often like $10. And the difference in performance can be negative. I would recommend running parts at the manufacturers spec and not caring about whether the lowest frames you get are 75 vs 80, but that's just me. If you want the blingee lights and super fast 4k ram, have a blast. It's not particularly making a difference in a way that actually matters aside from psychologically.
|
|
#
?
May 26, 2019 20:03
|
|
|
Broken Machine posted:Here is one of myriad benchmarks showing it doesn't matter. Oh wow, a 4% multicore increase, and otherwise flat performance. It's not worth buying bdie for most people or anyone really, which has a huge price premium of around an extra $150 right now, but it's certainly worth buying a mid-high range kit with tight timings and speeds if the price premium is minimal. It also depends what you're doing. Factorio and certain other simulation games will see massive performance improvements. In excess of 20%. Other games that aren't hammering memory will see hardly any, although on zen1 and zen+ there will always be some degree of performance increase due to the architecture. Khorne fucked around with this message at 20:09 on May 26, 2019 |
|
#
?
May 26, 2019 20:06
|
|
|
Paul MaudDib posted:So 150W for a 8C at 4.5 GHz? Power considerations were definitely something that edged me towards AMD. My old i5-3570K's TDP was 77W. The Ryzen 7 2700's is 65W. I'm getting 4 more cores and 12 more threads out of it for less power.
|
|
#
?
May 26, 2019 20:08
|
|
|
Khorne posted:2x16 of DDR4 3000 C16 is $129-$139.99 right now. The slower stuff costs about the same if not more. Absolutely, buy the least expensive reputable ram that's available at the fastest speed. Paying an extra $100 a stick for massively overclocked ram is silly, and if you happen to pick up 2400 on sale you're not hampering your system.
|
|
#
?
May 26, 2019 20:08
|
|
|
Also I think there might be some misunderstanding about latency here, most 3000 and 3200 kits have CAS latency 16 or better, equivalent to 12 or 13 at 2400. There are almost no kits at 2400 with lower latency than that.
|
|
#
?
May 26, 2019 20:09
|
|
|
Khorne posted:2x16 of DDR4 3000 C16 is $129-$139.99 right now. The slower stuff costs about the same if not more. 3000 C16 is the slower stuff. 3000 C15/3200 C16 is not that expensive. But yeah the price difference between 3000C16 and, say, 2400 C17 is like zero. Ten bucks over 16GB or something. Broken Machine posted:Absolutely, buy the least expensive reputable ram that's available at the fastest speed. Paying an extra $100 a stick for massively overclocked ram is silly, and if you happen to pick up 2400 on sale you're not hampering your system. Yeah if you buy 2400 in 2019 you're hampering your system. Even on Intel. Especially on Intel, maybe. iospace posted:Power considerations were definitely something that edged me towards AMD. You might want to measure that... the 2000 series Ryzens don't really respect TDP limits. Older intels do, newer Intels and newer Ryzens have started pretending that TDP is measured at base clocks and pushing down the base clocks to cover the growth in core count. Paul MaudDib fucked around with this message at 20:13 on May 26, 2019 |
|
#
?
May 26, 2019 20:10
|
|
|
This is one of my most favorite articles. RAM speed matters, cost is of course a huge competing factor, but the difference is pretty significant. https://www.eurogamer.net/articles/digitalfoundry-2016-is-it-finally-time-to-upgrade-your-core-i5-2500k
|
|
#
?
May 26, 2019 20:13
|
|
|
PS: spending an extra $100 a stick on the stupid 4700 kits may be actively counterproductive since XMP will then cue a much higher system agent voltage that is probably going to murder your processor within like six months or something. You really want a kit that is specced to run at 1.35V or maybe 1.4V, the 1.5V kits are LN2 overclocker territory for those people who will kill a processor in a week or something, so the SA voltage is not a problem. Thus spaketh Buildzoid. https://www.youtube.com/watch?v=HLNk0NNQQ8s That said, if you really want to waste money? Some 1.35V 4000 C19 is maybe another $70 over a base level kit ($35 a stick extra) and 4133 C17 is only about $270 per 16GB. If you're chasing CPU performance for high-refresh gaming then whatever, it's really the only place you can go upwards from a 9900K(S) at the moment.
|
|
#
?
May 26, 2019 20:18
|
|
|
Paul MaudDib posted:3000 C16 is the slower stuff. 3000 C15/3200 C16 is not that expensive. I bought bdie because I just don't care about price and want the extra frames, but if I had known that 3000 C16 price would crater from ~$220 for 32gb to $130 for 32gb I'd have not bought bdie at all because it's way overpriced. Khorne fucked around with this message at 20:29 on May 26, 2019 |
|
#
?
May 26, 2019 20:19
|
|
|
Paul MaudDib posted:I was serious earlier, if Lisa blueballs the crowd again she's going to get booed off the stage. If she does, that's fairly likely well, not literally, but she won't exactly receive applause HalloKitty fucked around with this message at 20:30 on May 26, 2019 |
|
#
?
May 26, 2019 20:22
|
|
|
|
| # ? Jun 7, 2024 00:16 |
|
|
Paul MaudDib posted:3000 C16 is the slower stuff. 3000 C15/3200 C16 is not that expensive. On Intel, the max recommended ram speed is still 2666. The difference between 2400 and 2666 is negligible. Aside from edge cases and overclocking it is essentially useless.
|
|
#
?
May 26, 2019 20:23
|
|