|
Paul MaudDib posted:Agreed on the smooth curves but it basically looks like a black Xbox 360 so what's the big deal? christ and i thought the ncase was difficult to work in
|
#
?
Dec 23, 2015 02:02
|
|
|
|
| # ? Jan 23, 2025 20:35 |
|
|
Malcolm XML posted:christ and i thought the ncase was difficult to work in You didn't work hard enough on a mITX build if your hands don't look like a cat mauled them.
|
|
#
?
Dec 23, 2015 02:23
|
|
|
GrizzlyCow posted:The etnews article is in English. Yes, but it's in what appears to be machine-translated English, or translated by someone who doesn't have a great grasp of English just yet. Still don't trust it 100%.
|
|
#
?
Dec 23, 2015 02:29
|
|
|
SwissArmyDruid posted:Today, in self-explanatory URLs: This makes sense from AMD's side, there's really no risk on their end; all of the risk is assumed by Samsung. There probably isn't any up-front cost on AMD's side, just guaranteed min order volume once production starts. If Zen is successful then AMD shouldn't have any problems meeting those order volumes. If it's not successful then Samsung can get in line with everyone else at bankruptcy court.
|
|
#
?
Dec 23, 2015 16:36
|
|
|
I could also see this deal coming about because 1) Samsung would be a great way for AMD to get into mobile, 2) AMD has GPU expertise that Samsung could use to get away from Mali. If AI and Zen are successful, there is no reason the companies wouldn't be rolling in money and continue to develop a stronger relationship. Wonder if AMD could also pull off something similar with Apple "Yea, just switch to Zen and you'll have the help of RTG on your GPU designs". EmpyreanFlux fucked around with this message at 18:23 on Dec 23, 2015 |
|
#
?
Dec 23, 2015 17:08
|
|
|
Samsung is also the only vendor in the world that can produce ASICs, interposers, and HBM, all together in-house. There is a benefit to having a one-stop shop, especially when co-integrating multiple vendors items at the package level is still largely voodoo.
|
|
#
?
Dec 23, 2015 18:12
|
|
|
I just built the first AMD machine I've built in years: AMD 9590 4.7 Ghz Octocore Asus ROG Crosshair V Formula Z Mainboard 10GB DDR3 (whatever I had laying around) Nvidia GeForce GTX 660 OC 2GB 300 GB SAS (OS) 4TB SATA (Storage) 1200 Watt PSU. Pretty happy with it for whatever I had laying around. And now he gets to sit next to my home server: Dell R815 4 x 12 Core AMD Opteron 2.2 Ghz 96 GB ECC DDR3 6 x 146 GB SAS in RAID-10 (OS/Applications/VMs) 8 x 1TB SATA in RAID-5E (Storage/Applications) And next to my old box: Dell R905 4 x 4 Core AMD Opteron 2.0 Ghz 128 GB ECC DDR2 6 x 146 GB SAS in RAID-10 CommieGIR fucked around with this message at 23:47 on Dec 23, 2015 |
|
#
?
Dec 23, 2015 23:44
|
|
|
CommieGIR posted:AMD 9590 4.7 Ghz Octocore Please tell me you didn't pay money for this chip.
|
|
#
?
Dec 24, 2015 00:18
|
|
|
pretty sweet opterons tho
Anime Schoolgirl fucked around with this message at 01:03 on Dec 24, 2015 |
|
#
?
Dec 24, 2015 00:43
|
|
|
CommieGIR posted:I just built the first AMD machine I've built in years: Ah. You're going to overclock it?
|
|
#
?
Dec 24, 2015 00:52
|
|
|
Boiled Water posted:Please tell me you didn't pay money for this chip. It was free. Client was going out of business, and they had hardware they purchased for their Marketing guy. GrizzlyCow posted:Ah. You're going to overclock it? I've looked up overclocks for it, and it seems like they barely push 5.0 Ghz out of it. I've got it water-cooled, so I'm not above taking a shot at it, but it performs pretty well as it.
|
|
#
?
Dec 24, 2015 01:58
|
|
|
CommieGIR posted:I've looked up overclocks for it, and it seems like they barely push 5.0 Ghz out of it. I've got it water-cooled, so I'm not above taking a shot at it, but it performs pretty well as it. You better have some rad watercooling on that thing; a 5 GHz Piledriver can cause a decent AIO watercooler to overload and anything too far above that is limited by your motherboard's power delivery. HardwareCanucks discovered this the crash-filled way..
|
|
#
?
Dec 24, 2015 03:51
|
|
|
Kazinsal posted:You better have some rad watercooling on that thing; a 5 GHz Piledriver can cause a decent AIO watercooler to overload and anything too far above that is limited by your motherboard's power delivery. HardwareCanucks discovered this the crash-filled way.. I'm not likely going to overclock it, and I have an Antec water cooler on it right now, it seems to handle the stock speed just fine.
|
|
#
?
Dec 24, 2015 03:55
|
|
|
It looks like it really does push about 250W. That's possible on a 120mm radiator - the 295x2 pushes 500W on the same cooler size. But man that's not gonna be fun. However - it'll perform somewhere between an i3 4170 and an i5-4460, just at 3x the power consumption. This has to be the one thread where we advise you not to buy the titular products. Except for maybe like a furry thread or something. It's mostly just a Athlon XP / X64 / X2 nostalgia thread with some Phenom X2 guys thrown in for good measure. The glory days. The Opterons sound cool though. Why don't you tell us about those. Can you put faster constructor cores there and like crunch videos or something? Paul MaudDib fucked around with this message at 04:38 on Dec 24, 2015 |
|
#
?
Dec 24, 2015 04:35
|
|
|
Paul MaudDib posted:This has to be the one thread where we advise you not to buy the titular products. Except for maybe like a furry thread or something. It's mostly just a Athlon XP / X64 / X2 nostalgia thread with some Phenom X2 guys thrown in for good measure. The glory days.  The 9590 Was free. So was the motherboard.The R815 is running 4 x 6174 Opterons, I use it for MATLAB and Simulations. It also runs Plex and hosts the controller for my Powervault Disk Array. http://www.cpu-world.com/CPUs/K10/AMD-Opteron%206174%20-%20OS6174WKTCEGO%20(OS6174WKTCEGOWOF).html And the AMD 9590 isn't THAT terrible: https://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-9590+Eight-Core&id=2014 Its between an i7 4770k and an i7 4790k. Nothing against the i7, I've got one in my Dell Precision. CommieGIR fucked around with this message at 05:20 on Dec 24, 2015 |
|
#
?
Dec 24, 2015 04:52
|
|
|
Bear in mind Passmark is pretty much perfect for that CPU - things are going to play out very differently for most actual usage.
|
|
#
?
Dec 24, 2015 05:49
|
|
|
CommieGIR posted:And the AMD 9590 isn't THAT terrible: http://cpu.userbenchmark.com/Compare/Intel-Core-i5-4690K-vs-AMD-FX-9590/2432vs1812 http://www.anandtech.com/bench/product/1261?vs=1289 On the other hand - nothing against AMD, I ran AMD processors since my K6-2 400 up until my Phenom II, but nowadays Intel just flat out wins. Intel ties in most productivity tasks except for highly threadable ones and the win is usually slight. And in gaming: Intel wrecks single threaded performance, Intel wins in average FPS, and the ownage gets even harder when you look at minimum FPS which are critically important to stutter free gameplay. Practically speaking the 9590 is somewhere between an i3-4170 and an i5-4460 in real world performance, just pulling 2-3x the power to do the job. The 4690K absolutely smokes it and AMD has no effective answer to the 4790K let alone the 6700K. Nowadays AMD processors that aren't based on Jaguar are just about unjustifiable vs the bargain priced LGA2011v3 processors. Jaguar/Kabini/etc is (barely) justifiable vs Baytrail-D but nothing past that is particularly notable anymore. Paul MaudDib fucked around with this message at 06:09 on Dec 24, 2015 |
|
#
?
Dec 24, 2015 06:02
|
|
|
CommieGIR posted:And the AMD 9590 isn't THAT terrible: That entirely depends on how you look at it:  Chowing down on power:  You can definitely cherry pick benchmarks that make it look almost competitive if you entirely ignore power draw, and ones that make it look awful if you concentrate on single-thread performance (as a lot of games still will benefit from). Of course, once the bottleneck generally shifts to the GPU, the difference is smaller; but you still have a space heater. All that said, gently caress it, it was free; so was the board as you mentioned, so you might as well play around with it. HalloKitty fucked around with this message at 10:54 on Dec 24, 2015 |
|
#
?
Dec 24, 2015 10:44
|
|
|
Pay more to do less, the AMD way.
|
|
#
?
Dec 24, 2015 16:30
|
|
|
I'm a little confused here. Why are we measuring FPS for the CPUs? go3 posted:Pay more to do less, the AMD way.  Or we could just have ourselves a monopoly.
|
|
#
?
Dec 24, 2015 16:34
|
|
|
CommieGIR posted:I'm a little confused here. Why are we measuring FPS for the CPUs? Why aren't we? IIRC that DX9 graph was populated using the same GPU setup with each of those processors. Higher FPS means the CPU is dealing with driver overhead and other issues more quickly, giving the GPU setup less downtime and smoother gameplay.
|
|
#
?
Dec 24, 2015 16:51
|
|
|
Intel has had a de facto monopoly on the performance desktop market since Sandybridge. Laptop-derived stuff like FM2 and AM1 are their only marketable products nowadays and only in niche applications. Measuring FPS in various games for various CPUs while holding the rest of the hardware constant is very straightforward. It's a real world benchmark - you're not buying the computer to play Linpack, right? Ideally it would include minimum fps or some other measure of microstutter too. At this point the FX series is questionable even if it's given to you. It's bested by a midrange processor that's 4 years old, and it consumes like 3x as much power to do so. I like AMD GPUs and I don't normally go for the efficiency argument but that's past what even I can stomach. 225w is a lot of loving power and heat. Single threaded performance is king in gaming, period. Also that's not even getting into overclocking - you can overclock any of the K processors like crazy and then they dunk on AMD even harder. Especially when you look at frame pacing / minimum FPS / frametimes. Paul MaudDib fucked around with this message at 17:01 on Dec 24, 2015 |
|
#
?
Dec 24, 2015 16:55
|
|
|
CommieGIR posted:I'm a little confused here. Why are we measuring FPS for the CPUs? To evaluate the relative performance between chips. The market of "enthusiast x86/64 processors" is not something we are going to make a difference in by buying inferior products.
|
|
#
?
Dec 24, 2015 17:00
|
|
|
CommieGIR posted:I'm a little confused here. Why are we measuring FPS for the CPUs? They're all likely using the same exact GPU and lowest settings to shift as much workload to the CPU as possible to measure CPU performance? TBF, If you can run a game on max settings the differences aren't as dramatic (from 30-40% down to 15-25%), but still present and you consume too much power for that performance. Also using that as a scale indicates that Zen will be roughly equal to slightly better than Ivy Bridge clock for clock while consuming less power. Hope they can overclock well and are cheap :\.
|
|
#
?
Dec 24, 2015 17:04
|
|
|
Jago posted:To evaluate the relative performance between chips. the implication here being that the CMT method also lost out to intel in high performance workspaces, and AMD's never had much of a foothold in laptops due to market inertia and probably never will unless they can outpace intel in both features and performance at the same price and power bracket poo poo's fairly rough
|
|
#
?
Dec 24, 2015 17:27
|
|
|
Paul MaudDib posted:Intel has had a de facto monopoly on the performance desktop market since Sandybridge. Laptop-derived stuff like FM2 and AM1 are their only marketable products nowadays and only in niche applications. Ah okay, that makes more sense.
|
|
#
?
Dec 24, 2015 20:22
|
|
|
HalloKitty posted:That entirely depends on how you look at it: How embarrasing for AMD that most of their current products are still not better than a processor they released 5 years ago. It was embarrasing when bulldozer first came out, but how hard is it to beat a 5 year old product when you have more than a full node advantage. E: even worse when you look at the APU's. AMD APU's: slightly better than a 6.5 year old quadcore NihilismNow fucked around with this message at 15:09 on Dec 26, 2015 |
|
#
?
Dec 26, 2015 15:02
|
|
|
I'm left wondering how a Deneb can outpreform a similarly clocked Thuban by any such margin.
|
|
#
?
Dec 26, 2015 18:35
|
|
|
GrizzlyCow posted:I'm left wondering how a Deneb can outpreform a similarly clocked Thuban by any such margin. It's the strictly higher clock rate doing that (3.7 vs 3.3 respectively), but the Thuban still has a slight advantage in frames per clock (27 vs 28 respectively). I still wonder how a 28nm Deneb/Thuban would perform. I don't think performance would much improve but they'd likely run really cold compared to Construction cores.
|
|
#
?
Dec 26, 2015 18:42
|
|
|
GrizzlyCow posted:I'm left wondering how a Deneb can outpreform a similarly clocked Thuban by any such margin. You mean x4 980 vs x6 1100t? 3.7 GHz is x4's base clock and x6's boost clock, thuban can't hold max boost.
|
|
#
?
Dec 26, 2015 19:02
|
|
|
Oops. Yeah, I mistook the X6 1100T's turbo clock for its base clock. It's a drat shame that outfits like TechReport doesn't include Phenom II results in their tests since I'd like to see how Deneb and Thuban holds up in modern games that do make use of more than 4 cores. Especially if they're OC'd to FX levels.
|
|
#
?
Dec 26, 2015 19:23
|
|
|
GrizzlyCow posted:Oops. Yeah, I mistook the X6 1100T's turbo clock for its base clock. A Thuban under water will beat a Piledriver clock for clock IIRC, let alone Bulldozer.
|
|
#
?
Dec 26, 2015 19:35
|
|
|
http://www.overclock3d.net/articles/cpu_mainboard/amd_have_revealed_their_new_wraith_cpu_stock_cooler/1quote:Let's hope that this change in CPU cooler design has not been prompted by how hot their new Zen CPU cores run as we certainly do not want another Volcano CPU from AMD, especially given how cool Intel CPUs run these days.
|
|
#
?
Jan 7, 2016 15:43
|
|
|
quote:The new, premium AMD Wraith Cooler features a near-silent noise profile quote:it is clear to hear how much quieter AMD's new reference cooler is. I'm OK with that.
|
|
#
?
Jan 7, 2016 16:07
|
|
|
When Intel is no longer shipping heatsinks with their performance parts, did this new heatsink really need to exist?
|
|
#
?
Jan 7, 2016 17:20
|
|
|
I'm not sure AMD is operating on the same consumer disregard tier as Intel though
|
|
#
?
Jan 7, 2016 17:22
|
|
|
Amd doesn't have performance parts to not ship heatsinks with.
|
|
#
?
Jan 7, 2016 17:22
|
|
|
Don Lapre posted:Amd doesn't have performance parts to not ship heatsinks with.
|
|
#
?
Jan 7, 2016 17:23
|
|
|
I run some FX-6300 (I think?) Six core thing in my home desktop. Winters get cold here in New England, so I can save on the heat by just running a bunch of VMs in the background. In the summer though, between the dust and the stock cooler, the thing has a tendency to hard-crash. Basically, pretty good if you want a space heater that can also do video encoding.
|
|
#
?
Jan 7, 2016 18:03
|
|
|
|
| # ? Jan 23, 2025 20:35 |
|
|
I don't consider Intel's new thing of not including a heatsink with K parts to be a bad thing. I do not know anyone who does not run an aftermarket cooler on K parts anyways, except when said aftermarket heatsink doesn't work, either because of DOA or age, and even then, only as a stopgap measure.
SwissArmyDruid fucked around with this message at 18:10 on Jan 7, 2016 |
|
#
?
Jan 7, 2016 18:08
|
|