|
MaxxBot posted:IIRC Zen 2 is supposed to have full speed AVX Intel will just bring out AVX-1024 to stay one step ahead. Never mind that the chip will have to clock down to 800MHz to avoid spontaneous combustion.
|
# ¿ Jul 22, 2017 10:35 |
|
|
# ¿ Apr 24, 2024 17:46 |
|
Pablo Bluth posted:They should probably introduce AVX512 too. Even if it's not that relevant to most people, it gives a poor impression to lose badly in subset of benchmarks. Munkeymon posted:The compression blur just makes it
|
# ¿ Sep 5, 2017 20:43 |
|
99% of people will never upgrade their laptop so it's a loss of performance for nearly all users.
|
# ¿ Oct 27, 2017 22:17 |
|
Vega eh? Looking forward to all the scalded junk lawsuits.
|
# ¿ Nov 6, 2017 21:46 |
|
Additionally: there's no reason to suppose any of them will work.
|
# ¿ Sep 21, 2018 01:13 |
|
Roughly how long is it after engineering samples become available before the retail release?
|
# ¿ Oct 2, 2018 21:41 |
|
Fair enough, expose the 95W TDP as being the lie we all knew it was. But hamstringing the chip then measuring performance doesn't seem like a very interesting thing with regard to making a purchasing decision.
|
# ¿ Nov 10, 2018 16:29 |
|
Lmao at anyone giving that amateur hour fabrication any credence whatsoever. It's bullshit and you all know it.
|
# ¿ Dec 3, 2018 22:23 |
|
Oh, well in that case Ryzen 4700X 8/16 6GHz boost* $299.99 *non-exotic cooling (480mm chiller)
|
# ¿ Dec 3, 2018 23:11 |
|
I, too, look forward to buying far more cores than I need. I'm glad I didn't get a 8700k when they came out. The jump from 4 to 6 cores looked great at the time, but kind of anaemic now.
|
# ¿ Jan 25, 2019 00:22 |
|
AMD absolutely should sell Zen 2 cheaper than the equivalent i7/9s. If they're the same price, people will just pick the one they consider a safer bet, and that's Intel. But I also agree that people who think they're getting more than 8 cores cheap are in for a disappointment.
|
# ¿ Jan 26, 2019 15:58 |
|
The TDPs are officially worthless at this stage. 3600X -> 3700X is a jump of 50% more cores and 200 MHz, but somehow only takes only 10W more power
|
# ¿ Mar 3, 2019 00:45 |
|
DDR5 is likely a 2020 thing at this stage, right?
|
# ¿ Mar 3, 2019 11:37 |
|
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 |
|
ItBreathes posted:I wonder if I could start selling audiophile woo to gamers? KillerPCIe
|
# ¿ Mar 23, 2019 15:31 |
|
Would it be safe to say that DDR5 won't be supported until Zen 2's successor?
|
# ¿ Apr 20, 2019 00:06 |
|
Paul MaudDib posted:That doesn't make sense. If you've got enough top-binned chiplets to service the desktop market with a high-clock (dual-die) 16C product as well as downmarket R5 products using single 8C dies, you've got enough top-binned chips to service the (much smaller) HEDT market as well. Given the tiny number of TRs sold, it might not be worth AMD's time to invest in them.
|
# ¿ May 15, 2019 20:04 |
|
Lookin' forward to the Zen2 benchmarks where the 9900k figures are measured with HT off to reflect real world use
|
# ¿ May 16, 2019 08:18 |
|
Munkeymon posted:Seems much more likely that they spent decades not thinking about security at the microcode level and were blindsided by recent advances. I read an article back a while back that said essentially that there's a well known and studied list of vulns which chip makers were expected to design around, but Intel didn't because of the performance cost. AMD did try, which is why it wasn't hit with as many of these issues.
|
# ¿ May 16, 2019 23:00 |
|
The 3900X is already $500, and they're adding a third more cores for the 3950X. I'd be surprised if it comes in under $700.
|
# ¿ Jun 9, 2019 20:42 |
|
Simple, Intel would go back to their traditional business of selling memory.
|
# ¿ Jun 15, 2019 11:50 |
|
What's the deal with Ryzen 3 APUs? All I could find was the 3400g, which despite the model number is 12nm. My company will be buying a batch of PCs soon and it would be great to get all dem sweet cores If there are no 7nm APUs coming within the next couple of months, what cheap options are there for adding basic (i.e. office level) graphics capability to an AMD system?
|
# ¿ Jun 20, 2019 21:16 |
|
Paul MaudDib posted:The absolute cheapest display adapters are still Tesla-based (GT210/GT730/etc) or Terascale-based (Radeon 5450/6450). If you want something with actual modern driver support I'd look at a RX 460, GT 1030, RX 570, or GTX 1050. If you want lots and lots of displayport outputs, probably a Quadro P400 or P600.
|
# ¿ Jun 21, 2019 07:44 |
|
We know Zen2 CPUs will cheaper for equivalent performance than Intel, but I got the impression that the motherboards will be pricier. Will new builds still end up cheaper by going AMD once RAM and motherboards are accounted for? I'm assuming that there will still be some motivation to buy better RAM for Ryzens.
|
# ¿ Jun 21, 2019 20:43 |
|
K8.0 posted:I'm curious how many of you planning to buy a 3900X have ever bought a $500 processor before. AMD seems to have a huge hype wave going in their favor right now. Last CPU I bought was a 4/4 3570k. The perceived wisdom of the time was to avoid overspending now because you're just going to replace it in 3 years with something far better. That didn't pan out, and it's pretty legit to spend more these days knowing that you're likely going to get 5+ years out of a decent CPU. A 12 core option is attractive because it should keep you ahead of the next gen consoles for their lifetime, so barring some exceptional IPC bumps in the meantime, it's a good investment. This is all assuming the 12/16 core AMD chips don't end up being shittier for games than the 8 cores (like Threadripper was) though. If they actually are worse, the 3900X becomes distinctly unappealing.
|
# ¿ Jun 30, 2019 17:26 |
|
Theris posted:12PM EDT/9AM PDT on the 7th, I think. This just covers up to the 3900x, right? I thought I heard the 3950x isn't out for a little longer.
|
# ¿ Jul 2, 2019 08:08 |
|
Wasn't Zen 2 supposed to be the last CPU that will work with the current socket?
|
# ¿ Jul 8, 2019 22:27 |
|
What a loving baby. I've never been tempted to watch one of his videos based on the rep he has. It's like the youtube equivalent of wccftech.
|
# ¿ Jul 23, 2019 20:29 |
|
TSMC's revenue graph from the article at the top of the page shows there's plenty of money in older processes. GoFlo's problem is they have to start investing in a new node eventually or they'll fade into embedded irrelevancy.
|
# ¿ Oct 20, 2019 11:36 |
|
TSMC are releasing 5nm Q1 next year so things are a little better than that.
|
# ¿ Nov 26, 2019 22:48 |
|
One factor keeping AMD from being more widely sold by Dell/HP/etc is that their APUs lag way behind their pure CPU releases. A 3700X might be great value but it becomes a lot less attractive once the cost of a dGPU is added on. Of course, I wouldn't be surprised if the big OEMs expect Intel to regain the top spot eventually, and they wouldn't want to sour the relationship over a relatively temporary state of affairs.
|
# ¿ Jan 4, 2020 23:38 |
|
What difference does recognition of technical improvements make?
|
# ¿ Jan 5, 2020 01:36 |
|
Does anyone have solid numbers on how quickly Zen2 cores throttle up when they transition from having nothing to do to being busy? Googling for p-state switching latencies didn't yield anything.
|
# ¿ Jan 11, 2020 16:14 |
|
No surprise, 8 channel TR just sounds like Epyc anyway. BangersInMyKnickers posted:Back in the first gen of Intel speedstep (long-rear end time ago) it was measuring p-state latencies in the nano-seconds. c-state transitions were in the microseconds, so still very low. The problems generally were in applications that created bursty workloads that would trigger rapid cycling between p or c states. The embedded power management controllers are considerably smarter than that now and know to leave things ramped up for a little longer to help mitigate against that specific problem. Perfect.
|
# ¿ Jan 14, 2020 00:24 |
|
Is there any info about when Zen3 is expected to be released? It's not such a big jump from Zen2 to 3 as it was from Zen+ to Zen2, right? So the gap between releases could potentially be a little less than the year-and-a-bit cadence of prior releases.
|
# ¿ Jan 26, 2020 23:16 |
|
Klyith posted:Soooooo AMD may be sitting back for a moment and holding the process improvement until next year, so they have something good to get people excited about AM5 or whatever the new platform for DDR5 is. If ryzen 4000 was substantially better than 3000, but then 5000 was a very minor improvement, that wouldn't work in their favor.
|
# ¿ Mar 7, 2020 00:09 |
|
X USB 3.60 S X USB 3.60 X X USB Series X
|
# ¿ Apr 1, 2020 09:05 |
|
Malcolm XML posted:Ddr is the last parallel single ended interface in wide use, they could switch to serial like IBM does.
|
# ¿ Apr 6, 2020 13:48 |
|
mdxi posted:All we know, in the form of actual AMD statements, is So assuming a small frequency bump from the refined 7nm process, we're talking maybe 15% faster than Zen2. Not bad, but it will be interesting to see if they can improve the cache latency to a significant degree.
|
# ¿ Apr 8, 2020 22:27 |
|
|
# ¿ Apr 24, 2024 17:46 |
|
Samsung's 3nm has been pushed out a few months too.
|
# ¿ Apr 15, 2020 09:39 |