|
FuturePastNow posted:nothing really beats a tower heatsink with a 120mm fan. Besides the things that beat a tower heatsink with a 120mm fan, such as a tower heatsink with a 140mm fan.
|
#
?
Jun 14, 2016 21:10
|
|
|
|
| # ? Apr 19, 2024 00:14 |
|
|
Water cooling is still technically air cooling, so you should just get an H100 
|
|
#
?
Jun 14, 2016 21:12
|
|
|
https://www.youtube.com/watch?v=-IyrR8KsSfQ So, at the very least, they've got working cpu, chipset and motherboard, which is looking promising as far as the launch schedule goes.
|
|
#
?
Jun 15, 2016 16:17
|
|
|
Maybe its a sign.. Doom and Zen = Demon spawn when you use Zen and Polaris together? Either way, AMD does really seem to like Doom as the demo game right now. Hmm.
|
|
#
?
Jun 15, 2016 18:41
|
|
|
EdEddnEddy posted:Maybe its a sign.. It's new, and has a big reconcilable name. Doom3 also was pretty demanding and forced a bunch of people to update (just to be disappointed) I remember getting horrible FPS on my Gefore4MX440SE in DOOM3.
|
|
#
?
Jun 15, 2016 18:45
|
|
|
pixaal posted:It's new, and has a big reconcilable name. Doom3 also was pretty demanding and forced a bunch of people to update (just to be disappointed) I remember getting horrible FPS on my Gefore4MX440SE in DOOM3. To be fair, Geforce 4 MX was poo poo at almost anything; (it was based on Geforce 2 with a new name, it wasn't related to the Geforce 4, and was worse than a Geforce 3) and every card with "SE" after is cut down even further.
|
|
#
?
Jun 15, 2016 18:49
|
|
|
HalloKitty posted:To be fair, Geforce 4 MX was poo poo at almost anything; (it was based on Geforce 2 with a new name, it wasn't related to the Geforce 4, and was worse than a Geforce 3) and every card with "SE" after is cut down even further. Yep. They were pretty crap. I should know; I had one in my crappy eMachines.
|
|
#
?
Jun 15, 2016 18:56
|
|
|
Hell even the Geforce 4 Ti 4600 (the top at the time) was getting its rear end handed to it by the ATI Radeon 9700 in Doom 3. Nvidia tried to get iD to demo it at E3 back then with Water Cooled 4600's but they turned them down for the 9700. There just was no comparison on how much better ATI was for Doom 3 at that time. I don't think Nvidia caught back up to ATI then until the 6000 series launched. The 6600's In SLI were pretty bad rear end from a lot of people I knew running it back then.
|
|
#
?
Jun 15, 2016 19:14
|
|
|
Just want to defend my mx440 cards honour by saying it was the best overclocking card I ever had, I could even play grand theft auto 3 from something other than the top down view. It went on to serve as my mum's card for a good ten years afterwards, eventually meeting it's demise when several capacitors started leaking
|
|
#
?
Jun 16, 2016 12:43
|
|
|
EdEddnEddy posted:Hell even the Geforce 4 Ti 4600 (the top at the time) was getting its rear end handed to it by the ATI Radeon 9700 in Doom 3. The 9700/9800 was a monster. I had a 9600 back then and it played Doom 3 and everything else pretty acceptably at 1024x768. I didn't really recall how the GeForce 4 performed back then (though my friend had one at the time, IIRC) so I looked it up.  Not quitte the 4600 but it wouldn't be 3 times faster anyway, would it? Also I guess ATI's horseshit OpenGL drivers didn't go back that far!
|
|
#
?
Jun 16, 2016 13:18
|
|
|
mobby_6kl posted:The 9700/9800 was a monster. I had a 9600 back then and it played Doom 3 and everything else pretty acceptably at 1024x768. I didn't really recall how the GeForce 4 performed back then (though my friend had one at the time, IIRC) so I looked it up. Aren't we all nostalgic for the time when ATI had the best video cards and AMD had the best CPUs. One day...~
|
|
#
?
Jun 16, 2016 13:43
|
|
|
And then, the 9600XT still wasn't really that great with its 4 ROP's once you saw what 8 ROP's on the 9700/9800 could do. It was crazy the jump ATI did after the Geforce 4. I remember the Geforce 3 Ti 500 being a freaking good king, and then the 4 Ti 4200 being actually a bit faster than that (And what I got in a warranty swap from Comp USA back then) but man the jump to the 9800 Pro and P4 3.0 back then from a P4 1.8 (though with RDRAM which was fast as hell) and a Ti 4600 was huge. I had a bad taste in my mouth from ATI drivers prior with cards like the Radeon DDR that could not render Giants Citizens Kabuto worth a drat even though it ran faster than my GF2MX, but the 9800Pro showed that once they got the Drivers and the Hardware, ATI was drat fine finally. Then the jump to that 7800GS was another big jump with 16ROPs for use on a 1680X1050 screen which the 9800pro couldn't really handle at that point.
|
|
#
?
Jun 16, 2016 18:25
|
|
|
I'm still waiting on Zen for an upgrade from a q9400 
|
|
#
?
Jun 16, 2016 20:19
|
|
|
Citizen kabuto ran fine on a Voodoo 2 so it is really weird the Radeon DDR couldn't run it very well.
|
|
#
?
Jun 17, 2016 19:10
|
|
|
NihilismNow posted:Citizen kabuto ran fine on a Voodoo 2 so it is really weird the Radeon DDR couldn't run it very well. It ran it really well, but the drivers didn't render stuff properly. Your gun as the mech guys would be see through as it couldn't render the sort of shiny/reflective effect properly. Stuff like that.
|
|
#
?
Jun 17, 2016 19:28
|
|
|
glide 
|
|
#
?
Jun 17, 2016 19:30
|
|
|
NihilismNow posted:Citizen kabuto ran fine on a Voodoo 2 so it is really weird the Radeon DDR couldn't run it very well. Hoo boy if you think ATI/AMD drivers are mediocre now, then understand that the sheer broken-ness of their drivers pre-R300 is something else.
|
|
#
?
Jun 17, 2016 19:32
|
|
|
SwissCM posted:Hoo boy if you think ATI/AMD drivers are mediocre now, then understand that the sheer broken-ness of their drivers pre-R300 is something else. Let's be fair, every driver was loving insane back then.
|
|
#
?
Jun 17, 2016 22:17
|
|
|
https://twitter.com/OC3D/status/743920227952529408 Interesting. Wish I had bought some back when it was less, but might want to jump on before Zen.. Hmm
|
|
#
?
Jun 17, 2016 22:38
|
|
|
Keep in mind this is more recovery to normal than it is AMD excelling.
|
|
#
?
Jun 17, 2016 22:53
|
|
|
fishmech posted:Let's be fair, every driver was loving insane back then. Nvidia were alright post-Riva128 and 3dfx's stuff was quite stable. Everything else was pretty poo poo.
|
|
#
?
Jun 18, 2016 08:39
|
|
|
Trying to do some product math here, though I would appreciate it if you'd check my work. I am actually in somewhat dire need of an upgrade on my daily driver. (Combo work-play CAD laptop) Zen isn't expected to be out until Q1 2017, although CES is in January, and if there's a time that AMD would like to have their best foot forward with mobile Zen, it *would* be CES. So HP, Dell, Lenovo, and so forth all show off their new laptop designs there. The question is though, is Zen going to launch 2017Q1, or is it going to start sampling to partners 2017Q1? Because if it starts sampling then, forget it, there's no way that Zen makes it into anything before they start releasing new products during the mid-April-to-early-September period. But there has been a rumor that Zen might have its release pushed up to 2016Q4. Which, if we know AMD like we do, effectively means "early 2017Q1" anyways. Whether or not you can buy a laptop with an AMD CPU or APU in the next 365 days depends entirely on during which of the three months of Q4 that Zen launches, paper or otherwise. In other words, if I was expecting to get a 14nm mobile GPU, 14nm workstation GPU of any kind, or AMD CPU/APU, I should just build the drat computer I need now and sell off the workstation laptop, because one year is way too long to wait and by then the value of this workstation laptop is going to crater by then. ...carry the one... you know, I'm not entirely certain that we're not all just secretly masochists. A year for mobile AMD products worth a poo poo? Jesus.
|
|
#
?
Jun 27, 2016 10:56
|
|
|
SwissArmyDruid posted:In other words, if I was expecting to get a 14nm mobile GPU, 14nm workstation GPU of any kind, or AMD CPU/APU, I should just build the drat computer I need now and sell off the workstation laptop, because one year is way too long to wait and by then the value of this workstation laptop is going to crater by then. This is always the answer: build the computer when you need it, don't count on hardware's performance until you can read reviews, and desktops are better value than laptops.
|
|
#
?
Jun 27, 2016 14:13
|
|
|
You do pay a really hefty penalty for cramming workstation performance in to a laptop, for sure, although they do tend to have decent-ish resale value. If you can get away with a desktop, build one now and just be done with it. (If you have to have a laptop though, I grabbed a Dell Precision 7510 a couple months back that has been excellent).
|
|
#
?
Jun 27, 2016 16:44
|
|
|
AMD IS PROFITABLE AGAIN! http://ir.amd.com/phoenix.zhtml?c=74093&p=irol-newsarticle&ID=2187336 So soon after moving to a new process across the board, no less.   ALSO, Zen Engineering samples spotted? Salt now so you're not salty later. http://www.guru3d.com/news-story/ams-zen-engineering-sample-specs-leaked.html
|
|
#
?
Jul 22, 2016 14:38
|
|
|
SwissArmyDruid posted:AMD IS PROFITABLE AGAIN! http://ir.amd.com/phoenix.zhtml?c=74093&p=irol-newsarticle&ID=2187336 Two separate L3 caches, hope that works out for them without any weird scaling as things go from 4 to 5 threads.
|
|
#
?
Jul 22, 2016 15:15
|
|
|
Xeons have done fine with completely separated cache banks. I'd hope that AMD has the same workarounds for that sort of thing that Intel does.
|
|
#
?
Jul 22, 2016 15:16
|
|
|
I want to be psyched by the 4 core 8 core chips that don't use hundreds of watts.
|
|
#
?
Jul 22, 2016 15:44
|
|
|
Anime Schoolgirl posted:Xeons have done fine with completely separated cache banks. I'd hope that AMD has the same workarounds for that sort of thing that Intel does. Show of hands, based on AMD's track record, who here trusts them to rip off what works while discarding what doesn't as a platform for future advancement in favor of trying to out-innovate from the bottom up? *doesn't raise hand* If AMD has proceeded to try to create something new in favor of letting Intel do their concept testing for them (hello, you guys are loving sharing patents, loving DO SOMETHING WITH THAT poo poo) I am not optimistic about Zen's chances. SwissArmyDruid fucked around with this message at 16:00 on Jul 22, 2016 |
|
#
?
Jul 22, 2016 15:54
|
|
|
That 2.8 - 3.2 GHz clock speed, even on engineering samples, is not encouraging.
|
|
#
?
Jul 23, 2016 04:29
|
|
|
sincx posted:That 2.8 - 3.2 GHz clock speed, even on engineering samples, is not encouraging. The issue seems to stem from GloFo, look at Polaris. Even so, Bulldozer was 2.6Ghz in ES, so I'm not discouraged because 3.2Ghz/95W/8C/16T is kind of actually ahead of Broadwell perf/watt it seems. Especially for A0 ES. Theory seems to be that AMD will target 4C/4T and 8T for 200$ price point, 6C/6T or 12T for 300$, and 8C/8T or 16T for 400$ bracket; as far as we know, AMD will not be offering dual cores at all, not even for APUs or mobile. Remember they're Haswell performance mostly so I think these prices make sense as long as clock speeds are 3.2-4.0Ghz. If they offer Snowy Owl and Naples for $800 and $1200 respectively that'd be a hell of an impact on the market.
|
|
#
?
Jul 23, 2016 06:27
|
|
|
FaustianQ posted:The issue seems to stem from GloFo, look at Polaris. Even so, Bulldozer was 2.6Ghz in ES It looks like Bulldozer's only 95W launch SKU, the FX-8100, shipped commercial chips at 2.8GHz. 
|
|
#
?
Jul 23, 2016 06:37
|
|
|
Twerk from Home posted:It looks like Bulldozer's only 95W launch SKU, the FX-8100, shipped commercial chips at 2.8GHz. Turbo was 3.1Ghz for all cores and 3.7Ghz for half or less. That'd mean 3.0Ghz baseclock, 3.33Ghz for 8C/16T and 4.0Ghz for 4C/9T or 8C/8T for Zen, let alone what it's 65W 4C/8T counterpart could do. That's...not exactly apocalyptic honestly, that seems pretty good. And this is working on the idea that GloFo hosed up 14nm just like 32nm.
|
|
#
?
Jul 23, 2016 06:59
|
|
|
sincx posted:That 2.8 - 3.2 GHz clock speed, even on engineering samples, is not encouraging. Given that if I want to get a 6-core Xeon for about $200 or so, I might be looking at base clocks of _1.6 Ghz_, this is incredibly encouraging. Especially for ES A0. https://www.amazon.com/Intel-Server-E5-2603V3-LGA2011-3-BX80644E52603V3/dp/B00M1BUY7K/ref=sr_1_3?ie=UTF8&qid=1468826910&sr=8-3&keywords=xeon FaustianQ posted:The issue seems to stem from GloFo, look at Polaris. Even so, Bulldozer was 2.6Ghz in ES, so I'm not discouraged because 3.2Ghz/95W/8C/16T is kind of actually ahead of Broadwell perf/watt it seems. Especially for A0 ES. To elaborate: Lets take that 8C/16T price point of $400. At that price point, the BEST Intel processor you can get is... an E5-2620 v4 (A Broadwell part, fyi). Which has a base clock of 2.1GHz and max turbo of 3.0GHz, for an MSRP of $417. Better max turbo with a massively better base clock? For ES A0? Yesplz. Remember, the more cores a part has, the lower the clocks tend to get, and if you want the kind of higher clocks you'd get on your enthusiast-grade quadcores, you're looking at shelling out an arm, leg, and kidney. http://ark.intel.com/search/advanced?s=t&FilterCurrentProducts=true&CoreCountMin=8&CoreCountMax=8&ThreadCountMin=16&ThreadCountMax=16 http://ark.intel.com/products/92986/Intel-Xeon-Processor-E5-2620-v4-20M-Cache-2_10-GHz Of course, by that same token, Zen could have poo poo IPC and those clocks could mean absolute jack poo poo. But if they're shipping engineering samples around this early, I take it as a good sign. Even if I *am* going to continue to dump on them by not believing they can follow through on what they're promising. SwissArmyDruid fucked around with this message at 08:06 on Jul 23, 2016 |
|
#
?
Jul 23, 2016 07:58
|
|
|
I, too, remember Bulldozer's utter failure to live up to hype and am prepared for the possibility than Zen will suck and AMD's engineers will become homeless. Lookin good so far, though.
|
|
#
?
Jul 23, 2016 20:52
|
|
|
FuturePastNow posted:I, too, remember Bulldozer's utter failure to live up to hype and am prepared for the possibility than Zen will suck and AMD's engineers will become homeless. The news at least sounds good so far unlike last time when all you had was JFAMD running his mouth all over the internet
|
|
#
?
Jul 23, 2016 21:28
|
|
|
http://www.pcper.com/news/Graphics-Cards/AMD-Introduces-Radeon-Pro-SSG-Professional-GPU-Paired-Low-Latency-Flash-Storage SSD + GPU = the ability to realtime scrub an 8K timeline at 90~ FPS. SwissArmyDruid posted:Show of hands, based on AMD's track record, who here trusts them to rip off what works while discarding what doesn't as a platform for future advancement in favor of trying to out-innovate from the bottom up? It's impressive, but I feel weirdly prophetic about this poo poo. I mean, this is _not nothing_. This is actually kind of incredible. At the very least, if AMD goes belly up, it's almost certain that RTG will live on. SwissArmyDruid fucked around with this message at 01:34 on Jul 27, 2016 |
|
#
?
Jul 27, 2016 01:31
|
|
|
The amount of mad some of the comments in that article have:quote:With people willing to pay $1000 for a single core Celeron if it comes with Intel's logo on it, I guess AMD had to make them feel a little better by adopting Intel's color. It will also be easier for the engineers to convince their clueless superiors, that the cards are Intel hardware and not AMD. "Look at the color. It is Intel's Blue".
|
|
#
?
Jul 27, 2016 03:32
|
|
|
SwissArmyDruid posted:http://www.pcper.com/news/Graphics-Cards/AMD-Introduces-Radeon-Pro-SSG-Professional-GPU-Paired-Low-Latency-Flash-Storage If anything, AMD now has a route to sell off any remaining Fiji dies at a very high margin. Clearly Fiji's home is compute heavy tasks and I'm sure video editors watching the presentation were drooling while trying to think were they can get $12,000 from.
|
|
#
?
Jul 27, 2016 11:04
|
|
|
|
| # ? Apr 19, 2024 00:14 |
|
|
Oil and gas is will be on top of this also.
|
|
#
?
Jul 27, 2016 16:08
|
|
