|
DrDork posted:Intel's stock getting crushed in July was 100% due to their own admission that 7nm would be delayed, and has shown basically no movement due to any AMD announcements before or since then. Yeah — INTC will be fine long term and it's not a bad time to buy it. The last time (A64 days), they bribed / paid / whatever'd to keep the OEM relationships strong and used the asset AMD didn't have (mountains of cash) to do what they needed. Not that it makes as much money as CPUs, but remember Intel makes flash (that's where all those n-2, n-3, n-4 fabs go...), networking adapters, FPGAs, is still a world-class foundry (though, they likely don't have PDKs/etc in a ready to go form for external houses like TSMC does) that is on-shore.
|
#
?
Oct 9, 2020 18:25
|
|
|
|
| # ? May 30, 2024 22:10 |
|
|
INTC is a fine long term play but where's the fun in that
|
|
#
?
Oct 9, 2020 18:31
|
|
|
Malcolm XML posted:INTC is a fine long term play but where's the fun in that I went to go look at the option chain for funsies and yep, there are definitely Jan 2023 calls available... pricey.
|
|
#
?
Oct 9, 2020 18:35
|
|
|
Don't buy stock because you think being an enthusiast in that area gives you a leg up
|
|
#
?
Oct 9, 2020 18:36
|
|
|
DrDork posted:I mean, don't get me wrong, Intel is fighting a losing battle right now, but it's a battle they're losing slowly enough that they've got plenty of opportunity to un-gently caress themselves. Yeah, Intel’s battle is really against themselves. AMD is just taking advantage of them losing it for now. Not that AMD isn’t doing great work on the CPU side, but if Intel were even doing ok work it would keep AMD in their place.
|
|
#
?
Oct 9, 2020 19:01
|
|
|
movax posted:
PLX was absorbed but really left alone. Broadcom was smart enough to realize the folks paying nearly 100 bucks for silicon in bulk have them specced into long term products so jacking up the supply chain for short term gain would be...shortsighted. If anyone still made a reliable PCI to ISA bridge, I know some folks who would love to buy them lol. You shouldn't spec a Microchip PCIe switch over a PLX especially at Gen4 speeds, unless you hate yourself or are penching pennies and at that rate, why are you doing a high end design? e: your previous work wouldn't have been with one of the early players in the all-flash market that failed when 08-12 recession happened would you? Crunchy Black fucked around with this message at 19:16 on Oct 9, 2020 |
|
#
?
Oct 9, 2020 19:09
|
|
|
Crunchy Black posted:PLX was absorbed but really left alone. Broadcom was smart enough to realize the folks paying nearly 100 bucks for silicon in bulk have them specced into long term products so jacking up the supply chain for short term gain would be...shortsighted. If anyone still made a reliable PCI to ISA bridge, I know some folks who would love to buy them lol. Nope — we built high-performance compute / HITL simulators that glued CPUs + FPGAs + friends together over various backplane species; I used the 8696 in a few designs where we had stupid amounts of fanout across the backplane and then a few years later was looking at the 8700-series for some spacecraft applications. I went to a training once for the Microchip switches — I guess it was Microsemi before they got bought, but either way they seemed to have a long ways to go before getting to PLX's level.
|
|
#
?
Oct 9, 2020 19:17
|
|
|
Sent you a PM.
|
|
#
?
Oct 9, 2020 19:29
|
|
|
Subjunctive posted:Yeah, Intel’s battle is really against themselves. AMD is just taking advantage of them losing it for now. Not that AMD isn’t doing great work on the CPU side, but if Intel were even doing ok work it would keep AMD in their place. To me Intel (and AMDs) long term battle is against ARM, not each other. It'll be interesting to see if the new ARM Macs actually start a real transition to ARM based desktop and laptops. Obviously that sort of thing would take a long time in the Windows world with all of those legacy applications people don't want to lose.
|
|
#
?
Oct 9, 2020 19:57
|
|
|
CFox posted:To me Intel (and AMDs) long term battle is against ARM, not each other. It'll be interesting to see if the new ARM Macs actually start a real transition to ARM based desktop and laptops. Obviously that sort of thing would take a long time in the Windows world with all of those legacy applications people don't want to lose. Yeah, legacy is never going away, at least not in our lifetimes. The question isn't "can ARM force people off legacy" (because the answer is gently caress no) but "can someone other than Apple write a x86 to ARM emulator fast enough that ARM can be competitive?"
|
|
#
?
Oct 9, 2020 20:05
|
|
|
No, not in any reasonable timeframe because the very reason Apple's ARM implementations are as good as they are have much more to do with their particular optimizations for them than the physical speed of the silicon. Unless Cavium fucks up and finds some sweetheart to give them a trillion dollars of series funding and becomes the default server chip moving forward but as has been pointed out that is probably a decade long transition period and then, legacy *still* exists.
|
|
#
?
Oct 9, 2020 20:11
|
|
|
CFox posted:To me Intel (and AMDs) long term battle is against ARM, not each other. It'll be interesting to see if the new ARM Macs actually start a real transition to ARM based desktop and laptops. Obviously that sort of thing would take a long time in the Windows world with all of those legacy applications people don't want to lose. wendell commented on this in his NVIDIA-ARM acquisition video, that Intel is now in the ironic position of having to rely on AMD to fend off ARM given their lack of progress over the last years. If all their customers bail to ARM in the meantime then it obviously becomes hard for Intel to win them back once they're back on their feet. Paul MaudDib fucked around with this message at 20:17 on Oct 9, 2020 |
|
#
?
Oct 9, 2020 20:11
|
|
|
It’s less x86 vs ARM and more expensive rear end x86 products vs big customers using something else or figuring out their own custom solutions
|
|
#
?
Oct 9, 2020 20:50
|
|
|
CFox posted:To me Intel (and AMDs) long term battle is against ARM, not each other. It'll be interesting to see if the new ARM Macs actually start a real transition to ARM based desktop and laptops. Obviously that sort of thing would take a long time in the Windows world with all of those legacy applications people don't want to lose. Apple will do what Apple does and few fucks will be given by anybody who doesn't use a Mac. I'll be amazed if the Adobe suite is somehow less of a total shitshow than it currently is when it's ported to ARM for MacOS. e: to be clear, they're great at controlling the entire widget and I'm sure they'll pull some magic off like 12-15hr battery life on their 85Whr notebooks but the catch will be you'll need that extra battery life to do poo poo that you could've otherwise done in less time on an Intel Mac. It's like the inverse of the PPC days when their CPUs were miniature furnaces, only now they've long binned off the good poo poo like Aperture that could compel people to stick with them through the transition. I'm sure Rosetta 2 will be functional, but it's still going to be less than ideal. Theophany fucked around with this message at 00:00 on Oct 10, 2020 |
|
#
?
Oct 9, 2020 23:57
|
|
|
ARM has always aimed at "really good compute in a small power envelope", which gives them incredible perf/watt but places a cap on absolute performance per core that's a lot lower than x86 chips. That's beginning to change though, with the Cortex X1 architecture, which is the first core that ARM has produced which is performance-oriented rather than power-optimized. Of course I'm not saying this is going to kill x86 on any timescale. I'm just saying that ARM might be more interesting (and less limited) than you previously thought. I'm really looking forward to seeing what they're capable of.
|
|
#
?
Oct 10, 2020 00:54
|
|
|
DrDork posted:Yeah, legacy is never going away, at least not in our lifetimes. The question isn't "can ARM force people off legacy" (because the answer is gently caress no) but "can someone other than Apple write a x86 to ARM emulator fast enough that ARM can be competitive?" If I could bet on anyone, it would probably be Nvidia.
|
|
#
?
Oct 10, 2020 01:09
|
|
|
Yeaahhhh, Nvidia and Apple are now all in with ARM and that is a pretty scary bellweather for the rest of the market imo x86 has a ton of momentum in the server world, no doubt, but I'm old enough to remember when the idea of x86 consumer parts muscling into big iron was itself considered a far off pipe dream. When people talk about x86 having too much momentum to be overtaken, I think of all those poor early codemonkeys that had to translate untold amounts of poo poo like bank software written for Alphas, SPARCs, Power, and Intel i860 (ironically, all RISC parts!), or even earlier stuff like 68000 CPUs. The explosive marketshare growth of x86 server should serve as a warning that poo poo CAN swap quickly with the right products and conditions. A lot of the x86 swap being viable was initially because the server designs (low volume, high price, bad R&D per sale ratio) were also sharing R&D and offsetting costs with home user stuff (high volume, low price, better R&D per sale ratio). That combined with increases in networking speed, workload balancing, and standardized rackmount form factors nuked the old big iron. Especially when the internet itself exploded in size and there was an insane need for piles of hardware that the old guard couldnt match. And frankly, there is a possibility we are looking at a similar set of circumstances with DPUs, GPUs/compute accelerators, and NVMEoF. Time will tell, but I'm pretty excited about the next few years.
|
|
#
?
Oct 10, 2020 01:52
|
|
|
The Apple transition kit leaked benchmarks show a 30% overhead for x86 emulation. Amazon graviton arm servers claim 40% better price performance than x86. With those figures running x86 on arm is already cheaper if you can live with lower clocks and buy more servers, and if you can run arm natively your saving a ton.
|
|
#
?
Oct 10, 2020 02:04
|
|
|
Perplx posted:The Apple transition kit leaked benchmarks show a 30% overhead for x86 emulation. Amazon graviton arm servers claim 40% better price performance than x86. With those figures running x86 on arm is already cheaper if you can live with lower clocks and buy more servers, and if you can run arm natively your saving a ton.
|
|
#
?
Oct 10, 2020 02:16
|
|
|
Kinda irrelevant, but I was just watching the PS5 teardown video that sony put out a few days ago. I was struck that this time around they specifically say "the CPU is an AMD Ryzen Zen 2", while for the PS4 they were like "it's x86 like a PC, but our own special design" and not mentioning AMD at all. I think it's a sign that AMD's brand being a plus is percolating out into the mainstream.
|
|
#
?
Oct 10, 2020 02:32
|
|
|
Klyith posted:Kinda irrelevant, but I was just watching the PS5 teardown video that sony put out a few days ago. I was struck that this time around they specifically say "the CPU is an AMD Ryzen Zen 2", while for the PS4 they were like "it's x86 like a PC, but our own special design" and not mentioning AMD at all. Well it was way before AMD's upturn from Ryzen lifting them out of the garbage and into relevancy again.
|
|
#
?
Oct 10, 2020 02:35
|
|
|
8-bit Miniboss posted:Well it was way before AMD's upturn from Ryzen lifting them out of the garbage and into relevancy again. The decision to go with AMD was obviously made years ago, yes. The decision to actually name-drop "AMD Ryzen Zen 2" in the teardown video they just put out was made much, much more recently. And, yeah, I think it is a good sign that AMD is now a reputable brand and not something that companies want to hide behind "x86-type processor" generalities.
|
|
#
?
Oct 10, 2020 02:49
|
|
|
DrDork posted:Intel's stock getting crushed in July was 100% due to their own admission that 7nm would be delayed, and has shown basically no movement due to any AMD announcements before or since then. As far as laptops go, if Tigerlake can't provide meaningful saturation across the price stack, it'll be Sky Lake cores vs Zen3 Cezanne and Sky Lake gets it's rear end kicked by Renoir as it is. Adding ~20% on top of that and it's hard to imagine any OEMs saying no to it, and I don't think customers will either. Alderlake is going to be Q3 2021, that's a really long span of time for Cezanne to be doing it's own thing. AMD's biggest issue in the laptop/server market has and will continue to be volume til the end of time, I think 35% laptop, 20% server is achievable but basically nothing more.
|
|
#
?
Oct 10, 2020 02:59
|
|
|
EmpyreanFlux posted:As far as laptops go, if Tigerlake can't provide meaningful saturation across the price stack, it'll be Sky Lake cores vs Zen3 Cezanne and Sky Lake gets it's rear end kicked by Renoir as it is. Adding ~20% on top of that and it's hard to imagine any OEMs saying no to it, and I don't think customers will either. Alderlake is going to be Q3 2021, that's a really long span of time for Cezanne to be doing it's own thing. Yeah, you're right that if Tigerlake 6c and 8c parts are not forthcoming, it's gonna be real hard for Intel to keep selling warmed over SkyLake forever. Volume is probably a big part of why AMD hasn't make a hard play for mobile yet--the margins are presumably low enough to not be attractive when they could be trying to put that wafer towards Epyc or TR or something higher-margin.
|
|
#
?
Oct 10, 2020 03:12
|
|
|
Josh Lyman posted:It’s not strictly performance. Isn’t power consumption a HUGE factor for commercial? Cooling. Its all around efficiently cooling everything and then making it as dense of as possible if you can cool an entire rack do you can put more stuff in
|
|
#
?
Oct 10, 2020 03:50
|
|
|
EmpyreanFlux posted:AMD's biggest issue in the laptop/server market has and will continue to be volume til the end of time, I think 35% laptop, 20% server is achievable but basically nothing more. I think if AMD isn't already looking at Samsung for a alternate fab than they should definitely start doing so soon because yeaaaah they're going to be very volume limited otherwise. Sammy's processes are looking to be behind TSMC's for a while yet but they're still a real step up from GF's already and at least they're interested in continued development too. If nothing else they could use Samsung to produce their APU's most of which don't need to be top clocked parts at all. Especially for the laptop market. I doubt that even with Samsung + TSMC that they'd be able to supply more than 50% of the x86 market but that would still be a big step up from their current status where the best they could do is a decent but still firmly minor chunk of the server, DIY, and laptop markets.
|
|
#
?
Oct 10, 2020 03:59
|
|
|
doomisland posted:Cooling. Its all around efficiently cooling everything and then making it as dense of as possible if you can cool an entire rack do you can put more stuff in Power matters too, because power -> heat but also some large cloud data centers are limited by how much power they can get delivered to the site. At a previous company the report you got after pushing something included the delta in "Wh per user per year" that you caused, and different teams had different budgets for it just like they did CPU time and storage.
|
|
#
?
Oct 10, 2020 04:02
|
|
|
PC LOAD LETTER posted:I think if AMD isn't already looking at Samsung for a alternate fab than they should definitely start doing so soon because yeaaaah they're going to be very volume limited otherwise. AMD ever getting to 50% of the x86 market would be success beyond their wildest dreams, even as that market changes into mostly server and laptop. I’d bet that the ~5 year trend is toward even more ARM client devices, so I guess the future of x86 is server only.
|
|
#
?
Oct 10, 2020 04:11
|
|
|
I’ve had some time to cool down and I’ve decided to wait a while to buy a 5900, if at all. Nothings really changed - the new AMDs catch up to the present Intels and exceed them slightly. So I’ll hang onto the 3700x and see how benchmarks and reviews handle gaming on a 3080 with the new crop of Zen3s if I really need to upgrade I can do it next year when it starts to matter.
|
|
#
?
Oct 10, 2020 04:57
|
|
|
Kraftwerk posted:I’ve had some time to cool down and I’ve decided to wait a while to buy a 5900, if at all. It seems to have been established there is going to be at least one more run of AM4 Ryzens in the next year or so, so waiting makes sense - that's what I'm doing.
|
|
#
?
Oct 10, 2020 05:02
|
|
|
WhyteRyce posted:Don't buy stock because you think being an enthusiast in that area gives you a leg up sure it does as long as u call urself an "analyst"
|
|
#
?
Oct 10, 2020 05:08
|
|
|
Twerk from Home posted:AMD ever getting to 50% of the x86 market would be success beyond their wildest dreams, even as that market changes into mostly server and laptop. I’d bet that the ~5 year trend is toward even more ARM client devices, so I guess the future of x86 is server only. Yeah but if they could ever actually get enough volume out the door AMD does have a product with Zen3 to have a shot at getting 50% or so. ARM clients in the form of laptops have been around since 2012 or so and they don't seem to be getting much traction. If anything the ARM guys seem to be really trying to target the server markets to get that sweet server cash rather than the much thinner margin client side of things.
|
|
#
?
Oct 10, 2020 06:39
|
|
|
AMD has also dabbled in Arm before. If Arm did make significant inroads into servers and desktops there is no reason why AMD wouldn’t make Arm processors as well.
|
|
#
?
Oct 10, 2020 06:43
|
|
|
AMD making ARM chips again means funneling money into Nvidia's pockets. They'd likely want to avoid that.
|
|
#
?
Oct 10, 2020 06:49
|
|
|
Business is business. Apple buys screens from Samsung. If there’s profit to be made they will do so.
|
|
#
?
Oct 10, 2020 06:50
|
|
|
Buying whole parts is different from investing resources developing upon an architecture that is owned by a direct competitor. Apple could switch suppliers in a single generation, licensing and creating an ARM processor requires far more effort, time and money. They're more likely to move towards RISCV, as a lot of the industry is probably doing with Nvidia absorbing ARM.
|
|
#
?
Oct 10, 2020 06:58
|
|
|
I'm not well-versed on the matter, but I think of ARM processors as being in mobile devices and such. Aren't pretty much all cellphones ARM? I know people have been talking about laptops and servers as possible ARM markets, but it already dominates (if not completely owns) the mobile device market. It seems like that would be where the money is. Competing with x86 seems like a possible avenue for the architecture but it seems like something of a secondary market.
|
|
#
?
Oct 10, 2020 08:12
|
|
|
What are my options for the nice motherboards for Ryzen 3?
|
|
#
?
Oct 10, 2020 08:16
|
|
|
Whitest Russian posted:What are my options for the nice motherboards for Ryzen 3? Depending on what your needs/wants are and how much you want to spent then you probably want B550 or X570. From there it's a matter of choice between brands and specific boards.
|
|
#
?
Oct 10, 2020 08:23
|
|
|
|
| # ? May 30, 2024 22:10 |
|
|
CaptainSarcastic posted:Depending on what your needs/wants are and how much you want to spent then you probably want B550 or X570. From there it's a matter of choice between brands and specific boards. I'd like to do some overclocking I guess and want to stay away from MSI/Gigabyte boards. Also ATX size. Whitest Russian fucked around with this message at 08:45 on Oct 10, 2020 |
|
#
?
Oct 10, 2020 08:27
|
|