|
Measly Twerp posted:I hope someone sticks an 8 core chip in a laptop. Would be nice to have a semi-portable workstation, even if it gets as bulky as some of those gaming laptops with dual 1080s inside. I doubt it would have to be as big as a 1080 GPU'ed one, you can already get those 7820HK's which are unlocked with a 1070 in a pretty slim chassis, so a 8 core that is unlocked but underclocked a bit should still work fine I'd imagine in something like that. Though if they let the laptop chips be unlocked as well, then yes they might need one of those SLI 1080 capable setups to keep the chip cook if you must OC it. I agree that having 8 cores for video editing on hand would be really awesome.
|
#
?
Mar 16, 2017 18:04
|
|
|
|
| # ? Oct 6, 2024 02:11 |
|
|
Pryor on Fire posted:lol still no gaming benchmarks on anandtech 2 weeks out. Sure hope someone at least got a nice beach house from that agreement! Digital Foundry also promised a review with frametime comparison videos, instead they posted "Switch Loading Times! Cartridges vs MicroSD vs Internal Storage!" an hour ago.
|
|
#
?
Mar 16, 2017 18:21
|
|
|
 sony vaio zSony were the only people with full blown 4 cores in 13" laptops and weren't afraid of being a bit fatter than a macbook (and thus didn't throttle to hell). Expensive as gently caress, but so worth it.
|
|
#
?
Mar 16, 2017 18:57
|
|
|
Truga posted:
Not to mention discrete NVIDIA GPUs and QUAD RAID 0 SSDs, still in a tiny, light chassis. So good.
|
|
#
?
Mar 16, 2017 19:06
|
|
|
Quad raid 0 sounds like you're just begging to lose all your data though.
|
|
#
?
Mar 16, 2017 19:08
|
|
|
fishmech posted:Quad raid 0 sounds like you're just begging to lose all your data though.
|
|
#
?
Mar 16, 2017 19:24
|
|
|
I think it was just dual raid0 SSD, but that came in later versions when SSDs were a less temperamental thing (and also supported trim, my first vaio didn't, let me tell you about how absolutely slow SSDs can get if they're nearly full and don't support trim  ).Said c2d vaio z still runs, too, my dad uses it now and it's better than any $800 shittop I could buy for him today. The cpu is a bit dated (but just fine for browsing/office poo poo), but it survived all kinds of mishaps, including falling off a server rack. There's a few scratches here and there, but it still works fine. The power socket feels a bit wobbly, but that's it. The perfect thing for old people to abuse and not break, lol
|
|
#
?
Mar 16, 2017 19:28
|
|
|
Risky Bisquick posted:They need to fix memory compatibility with their controller. Can they patch this with microcode or a bios update? As is it will create a problem when the more numerous R5 users can't get over 2400 with their 3000+ dual rank kits they already have. AMD has to individually test each kit and provide motherboard companies with the files to implement and then the motherboard companies have to update the bios.
|
|
#
?
Mar 16, 2017 19:29
|
|
|
Measly Twerp posted:I hope someone sticks an 8 core chip in a laptop. Would be nice to have a semi-portable workstation, even if it gets as bulky as some of those gaming laptops with dual 1080s inside. Yeah, and they'd still be pretty decently clocked to, like 2.0-2.2Ghz base and 2.8-3.2Ghz single core boost. My theory, based on the recently leaked road map, is AMD views Summit Ridge/Raven Ridge and Pinnacle Ridge as companion designs, one for a low power target and the other for raw performance.
|
|
#
?
Mar 16, 2017 19:33
|
|
|
I feel slightly differently. With how the desktop parts were the first to hit the market, followed by Naples in a few, and no Raven Ridge until next year, it seems like an inversion of the Intel binning process. Intel: Run Mobile and Server parts, take the crap-binned parts and put them into Desktop packages. AMD: Use the less-stingent Desktop market to sort out the problems, test batches, etc. Use this information to dial-in improvements to Server/Mobile parts. I feel this would also explain why the chips released thus far are so stinking cheap: We would essentially getting the non-semiconductor manufacturing equivalent of first-runs, setup parts, blemishes, and so forth. SwissArmyDruid fucked around with this message at 20:14 on Mar 16, 2017 |
|
#
?
Mar 16, 2017 20:05
|
|
|
Truga posted:I think it was just dual raid0 SSD They really did have quad raid 0 SSDs available, 256GB in total. Mine only has the dual raid 0 SSDs, although I did remove the optical drive and replace it with ANOTHER SSD, so I'm running triple raid 0 in mine, because hell, powerrrr! (Also, it's not like I use my laptop for anything super critical, so it wouldn't be that bad if something failed). But I think we're talking about different generations, mine has a Nehalem-based CPU in it, not Conroe or whatever flavour of C2D is in the one you mention.
|
|
#
?
Mar 16, 2017 20:16
|
|
|
Don Lapre posted:AMD has to individually test each kit and provide motherboard companies with the files to implement and then the motherboard companies have to update the bios. Per chipset?
|
|
#
?
Mar 16, 2017 20:36
|
|
|
Truga posted:AMD is absolutely better for basic media consumption and browsing or whatnot at same price points. Why? I've never run into any media or browsing situation where even an HD4000 became a problem and battery life is so poor on current AMD models that it isn't worth the tradeoff. Also HP does a full line of corporate models with AMD CPUs (the ProBooks and EliteBooks ending in 5). I'm not sure anyone ever bought one though because why would you?
|
|
#
?
Mar 16, 2017 20:43
|
|
|
This map was a little much for my iGPU. You generally have to try to find web sites like that, though.
|
|
#
?
Mar 16, 2017 20:52
|
|
|
That is hitting the CPU hard on my ancient dual core i7 but performance is fine. I'm not sure a current AMD APU would deal with it any better given how weak the CPU side is.
|
|
#
?
Mar 16, 2017 21:07
|
|
|
HalloKitty posted:They really did have quad raid 0 SSDs available, 256GB in total. Mine only has the dual raid 0 SSDs, although I did remove the optical drive and replace it with ANOTHER SSD, so I'm running triple raid 0 in mine, because hell, powerrrr! (Also, it's not like I use my laptop for anything super critical, so it wouldn't be that bad if something failed). I had 2 vaios, first a c2d one (that cost me 2000 euros after a 40% discount, yikes), then later an i7 one that was like 1300 refurb. The i7 one also still works, but I kinda trashed it avoiding a stray cat on the road. The monitor keeps developing dead pixels randomly, they travel (and at times disappear almost completely) around the display like some sort of worms. I tried fixing it, but nothing helps, the panel is just shot I think. I use it as a sort of backup now if people need something *now* for a few days, it works well enough for that. I don't need that much power these days, so I'm on "just" a chomebook pixel. That said, I'd starve for a month if it meant I could buy a good zen APU with 16-32 gigs of ram, ideally some of it HBM, and the pixel's display. Don Lapre posted:AMD has to individually test each kit and provide motherboard companies with the files to implement and then the motherboard companies have to update the bios. I thought bios was all on the cpu now? dissss posted:Why? I've never run into any media or browsing situation where even an HD4000 became a problem and battery life is so poor on current AMD models that it isn't worth the tradeoff. That's exactly the problem. poo poo battery isn't a function of amd or intel, it's a function of poo poo or good battery. My 2008 c2d vaio still lasts >5 hours on the battery I got on day 1, and it used to last 9, despite being a 25W chip. My current laptop has an dual core i7 that's 15W, and that's with a far more powerful igpu included so it saves a bunch there too, and it's still only 12h if I stretch it a bit. A decent battery absolutely makes all the difference. The issue is all AMD laptops are shitshows around the APU. Anyway, with an AMD APU, you could probably even do some light gaming pretty well, since they perform way better at normal resolutions than non-iris intel ones afaik. And I'd say simple games pretty much fit into "media and browsing" these days, what with html5.
|
|
#
?
Mar 16, 2017 21:09
|
|
|
Truga posted:That's exactly the problem. poo poo battery isn't a function of amd or intel, it's a function of poo poo or good battery. My 2008 c2d vaio still lasts >5 hours on the battery I got on day 1, and it used to last 9, despite being a 25W chip. My current laptop has an dual core i7 that's 15W, and that's with a far more powerful igpu included so it saves a bunch there too, and it's still only 12h if I stretch it a bit. A decent battery absolutely makes all the difference. The issue is all AMD laptops are shitshows around the APU. It kinda is though, the HP Elitebook 725 (AMD) is the same thing about an 820 (Intel) aside from the CPU and gets far, far worse battery life Truga posted:Anyway, with an AMD APU, you could probably even do some light gaming pretty well, since they perform way better at normal resolutions than non-iris intel ones afaik. And I'd say simple games pretty much fit into "media and browsing" these days, what with html5. The thing is I haven't run into an HTML5 games that won't work fine even on slow integrated.
|
|
#
?
Mar 16, 2017 21:16
|
|
|
Truga posted:Anyway, with an AMD APU, you could probably even do some light gaming pretty well, since they perform way better at normal resolutions than non-iris intel ones afaik. And I'd say simple games pretty much fit into "media and browsing" these days, what with html5. Is there any game out there that the laptop APUs could play acceptably that say, Intel Graphics 540 can't? I thought that the APU graphics advantage was pretty much gone.
|
|
#
?
Mar 16, 2017 21:24
|
|
|
My HD5500 chokes on anything 3d at 2560x1700. e: I honestly have no idea how 540 performs, I didn't even know there was a 500 series, heh. I can only presume it's still kinda crap, and hope zen apus with some HBM get released. e2: yeah apparently 500 series is about samey as amd apus, never mind then. Truga fucked around with this message at 21:31 on Mar 16, 2017 |
|
#
?
Mar 16, 2017 21:24
|
|
|
I got a Vaio SZ 645 something with only a 1.6Ghz C2D in it that's a 13.3" that has switchable graphics and an optical drive even for free from a client that was just about to throw out some old laptops. The stock 160G HDD was slow as crap and it only had 1G of ram, but upgrading it to it's max of 4GB and throwing in a 160G SSD made it into a perfectly serviceable little laptop even today. Also if anyone needs modern drivers for the special snowflake 8400GS for Windows 8/10, I was able to edit the last drivers Nvidia released for the 8000 series to work on this thing and it improved performance immensely over the last Win7 drivers that existed about 200.0 driver revisions back. The only thing(s) that stink on this thing is the battery (just too darn small but gets 2hrs~) the bad screen. (No such thing as "white" and 1280X800) and the wifi card is an Intel N chip that they stopped supporting and you have to run a older driver with or else it will BSOD anytime you try to actually push data through it. The older driver works fine except when you come out of sleep mode its 50/50 if it will work or not and need a reboot to clear up. Need to find a slightly newer replacement for it. Still, the fact that it has all this in such a small chassis way back when is impressive. A friend of ours had another version of an older SZ with a Centrino single core and a ATI 9700M that I played KOTOR on back in the day. It was amazing having that kind of power in a laptop back then. Too bad he had it stolen out of his Hotel room like literally a day after I fixed it all up for him.  ^^ Uncle has a Dell XPS 13 with a 520 in it. Thing faceplants on anything native res (4K~) and doesn't do much better even at 1080P. I have had some experience with the Iris 560 in the Skull Canyon NUC but even that struggles still with 1080P and more than medium detail.
|
|
#
?
Mar 16, 2017 21:29
|
|
|
Truga posted:My HD5500 chokes on anything 3d at 2560x1700. From notebookcheck, it looks like Intel 540 is a little faster than "AMD Radeon R7 M460", an entry level AMD dedicated GPU.
|
|
#
?
Mar 16, 2017 21:31
|
|
|
Twerk from Home posted:From notebookcheck, it looks like Intel 540 is a little faster than "AMD Radeon R7 M460", an entry level AMD dedicated GPU. I'd be surprised if it wasn't, that thing is saddled with DDR3. I really hope Raven Ridge is based off of Vega's feature set, TBR along with DDR4 and texture compression would go a long way towards helping with that memory bottleneck.
|
|
#
?
Mar 16, 2017 23:29
|
|
|
Wondering about the 4 core Ryzen, as they are cutting down each core complex, isn't this going to make the thread hopping problem significantly worse?
|
|
#
?
Mar 17, 2017 00:55
|
|
|
GRINDCORE MEGGIDO posted:Wondering about the 4 core Ryzen, as they are cutting down each core complex, isn't this going to make the thread hopping problem significantly worse?
|
|
#
?
Mar 17, 2017 01:38
|
|
|
I really am wondering how well the 4c stuff is going to run games, if the fabric's still tied to Ram speed. It's going to be even more dependant on the fabric bus and latency hit, with half as many cores per CCX, right? GRINDCORE MEGGIDO fucked around with this message at 08:38 on Mar 17, 2017 |
|
#
?
Mar 17, 2017 01:58
|
|
|
On Infinity Fabric Hexs and Quads are 3+3 and 2+2 deals (LMAO when Raven Ridge might be a better gaming platform) 
|
|
#
?
Mar 17, 2017 03:07
|
|
|
Also OrangeKhrush, who has been eerily on point for Ryzen, has posted on Anandtech that the inherent issues Ryzen is having with games has little to do at the software level and more that it's a stepping issue that's already been resolved in time for R5/HEDT/Server release. The best guess? The data fabric is no longer locked to IMC frequency in the new C stepping, but can be re enabled inside the BIOS. Why was this done in previous steppings? I guess some issue in quality of silicon?
|
|
#
?
Mar 17, 2017 03:24
|
|
|
FaustianQ posted:On Infinity Fabric So what if Inifab was 512bit would that help things?
|
|
#
?
Mar 17, 2017 03:43
|
|
|
wargames posted:So what if Inifab was 512bit would that help things? Yeah, but then it would be even harder to get good units and actually sell large quantities of this at a low price. Wide buses are hard.
|
|
#
?
Mar 17, 2017 03:49
|
|
|
So infinifab is cpu side and not motherboard, so we might see a 384bit bus toward the end of zen?
|
|
#
?
Mar 17, 2017 03:54
|
|
|
wargames posted:So what if Inifab was 512bit would that help things? Yeah that speed seems really low for connecting on-die CPU caches. An order of magnitude slower than GDDR5X memory? Math is hard at 4am 
|
|
#
?
Mar 17, 2017 04:00
|
|
|
FaustianQ posted:On Infinity Fabric Holy poo poo, I was totally convinced they would be single-CCX. I guess that's for gaming on their quads, I can only imagine some Intel manager spent the afternoon shouting into a phone to cancel the plans for a hyperthreaded i5
|
|
#
?
Mar 17, 2017 04:02
|
|
|
sauer kraut posted:Yeah that speed seems really low for connecting on-die CPU caches. An order of magnitude slower than GDDR5X memory? Math is hard at 4am It's not really, if the IMC and Infinity Fabric are run at 1:1 it'd be the same effective inter core communication speed as Intel's, so unless Intel is considered slow... Paul MaudDib posted:Holy poo poo, I was totally convinced they would be single-CCX. Depends on whether you get a B or C stepping apparently, LMAO at this level of silicon lottery.
|
|
#
?
Mar 17, 2017 04:33
|
|
|
Bless you AMD for opening preorders for Ryzen. I'd have bought launch silicon on day 1 if you hadn't.
|
|
#
?
Mar 17, 2017 04:35
|
|
|
So will the R5 1600X be the most up-to-date stepping?
|
|
#
?
Mar 17, 2017 05:04
|
|
|
SourKraut posted:So will the R5 1600X be the most up-to-date stepping? It's unknown, and it's only a rumor of C stepping, but it's coming from someone super reliable. And even if there is a C stepping, all indications are that it'll be mixed in with B steppings, because that went well that last time that happened.
|
|
#
?
Mar 17, 2017 05:12
|
|
|
If there is a C stepping, how fast is the fabric running? And is there a C stepping of Ryzen 5 and 7?
|
|
#
?
Mar 17, 2017 05:45
|
|
|
GRINDCORE MEGGIDO posted:If there is a C stepping, how fast is the fabric running? 3x108
|
|
#
?
Mar 17, 2017 05:54
|
|
GRINDCORE MEGGIDO posted:If there is a C stepping, how fast is the fabric running? Infinitely fast, duh.
|
|
|
#
?
Mar 17, 2017 07:03
|
|
|
|
| # ? Oct 6, 2024 02:11 |
|
|
But is it going to be fast enough for the HYPE TRAIN??
|
|
#
?
Mar 17, 2017 08:05
|
|