AMD Piledriver rumours ... and expert conjecture

Page 22 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Reynod

Reynod

Administrator
Mar 26, 2007
14,921
16
42,215
We have had several requests for a sticky on AMD's yet to be released Piledriver architecture ... so here it is.

I want to make a few things clear though.

Post a question relevant to the topic, or information about the topic, or it will be deleted.

Post any negative personal comments about another user ... and they will be deleted.

Post flame baiting comments about the blue, red and green team and they will be deleted.

Enjoy ...
 
Sort by date Sort by votes
Has the PHII x8 been released in EU/Asia?? Any idea on performance since this is what everyone complains that they should have done (PHII @ 32nm instead of BD)?
Click to expand...
no, they have never and will never make this now 🙁
to be fair to AMD this is very very difficult to get something this complicated right first time around im sure there are many fairly simple changes they can make, and the clockspeeds should increase as gloflo's 32nm gets more mature
 
Upvote 0 Downvote
I wonder what was wrong with it that AMD decided to go all BD (180w TDP??)?

Maybe they already had some idea of GF's problems with 32nm and decided not to risk starting another product line that they couldn't produce in volume.
 
Upvote 0 Downvote
much appreciated for the link..
Click to expand...

He's what he's reported on PPD.


I switched the 8120 over to Linux (Ubuntu 10.10). Everything is set to stock in the BIOS, however Linux reports the chip as running at 3.6GHz load (and the correct 1.4GHz at idle). I haven't yet gotten a bigadv (on any of my machines, however my i7 is still doing the awesome SMPs). The 8120 did however start a good SMP, a 7019, and is currently averaging 2:03 TPF. I checked the log and this seems to be accurate, and it is well over half way done. Furthermore with that TPF both Control and the SMP Calc agree : ~31000-32000 PPD!!! So of course the question s why? Without more WU's I can't be sure, but I never remember seeing that much of a speedup with my i7 back when I ran it in Ubuntu 10.10, even with special kernels. I guess it could also be the WU (7019). I guess we'll see as it finishes and gets more different WU's. Wow is this confusing and unpredictable, but mostly good so far.
Click to expand...


Here's a log from a recent 7010 in Linux, CPU was at 3.6GHz (probably), system consuming 200W, CPU temp (claimed) at about 40C. Control was claiming a hair over 27000 PPD:

11:50:15:Unit 00: Uploading 533.35KiB to 171.64.65.104
11:50:15:Connecting to 171.64.65.104:8080
11:50:15:Starting Unit 01
11:50:15:Running core: /var/lib/fahclient/cores/www.stanford.edu/~pande/Linux/AMD64/Core_a4.fah/FahCore_a4 -dir 01 -suffix 01 -lifeline 1043 -version 701 -checkpoint 15 -np 8
11:50:15:Started core on PID 3824
11:50:15:FahCore 0xa4 started
11:50:16:Unit 01:
11:50:16:Unit 01:*------------------------------*
11:50:16:Unit 01:Folding@Home Gromacs GB Core
11:50:16:Unit 01:Version 2.27 (Dec. 15, 2010)
11:50:16:Unit 01:
11:50:16:Unit 01😛reparing to commence simulation
11:50:16:Unit 01:- Looking at optimizations...
11:50:16:Unit 01:- Created dyn
11:50:16:Unit 01:- Files status OK
11:50:16:Unit 01:- Expanded 53448 -> 208564 (decompressed 390.2 percent)
11:50:16:Unit 01:Called DecompressByteArray: compressed_data_size=53448 data_size=208564, decompressed_data_size=208564 diff=0
11:50:16:Unit 01:- Digital signature verified
11:50:16:Unit 01:
11:50:16:Unit 01😛roject: 7010 (Run 1, Clone 157, Gen 63)
11:50:16:Unit 01:
11:50:16:Unit 01:Assembly optimizations on if available.
11:50:16:Unit 01:Entering M.D.
11:50:21:Unit 00: Upload complete
11:50:21:Server responded WORK_ACK (400)
11:50:21:Final credit estimate, 5267.00 points
11:50:21:Cleaning up Unit 00
11:50:21:Unit 01:Completed 0 out of 10000000 steps (0%)
11:52:34:Unit 01:Completed 100000 out of 10000000 steps (1%)
11:54:49:Unit 01:Completed 200000 out of 10000000 steps (2%)
11:57:04:Unit 01:Completed 300000 out of 10000000 steps (3%)
11:59:20:Unit 01:Completed 400000 out of 10000000 steps (4%)
12:01:35:Unit 01:Completed 500000 out of 10000000 steps (5%)
12:03:51:Unit 01:Completed 600000 out of 10000000 steps (6%)
12:06:07:Unit 01:Completed 700000 out of 10000000 steps (7%)
12:08:22:Unit 01:Completed 800000 out of 10000000 steps (8%)
12:10:38:Unit 01:Completed 900000 out of 10000000 steps (9%)
12:12:53:Unit 01:Completed 1000000 out of 10000000 steps (10%)
12:15:09:Unit 01:Completed 1100000 out of 10000000 steps (11%)
12:17:24:Unit 01:Completed 1200000 out of 10000000 steps (12%)
12:19:40:Unit 01:Completed 1300000 out of 10000000 steps (13%)
12:21:56:Unit 01:Completed 1400000 out of 10000000 steps (14%)
12:24:12:Unit 01:Completed 1500000 out of 10000000 steps (15%)
12:26:27:Unit 01:Completed 1600000 out of 10000000 steps (16%)
12:28:43:Unit 01:Completed 1700000 out of 10000000 steps (17%)
12:30:58:Unit 01:Completed 1800000 out of 10000000 steps (18%)
12:33:14:Unit 01:Completed 1900000 out of 10000000 steps (19%)
12:35:30:Unit 01:Completed 2000000 out of 10000000 steps (20%)
12:37:46:Unit 01:Completed 2100000 out of 10000000 steps (21%)
12:40:02:Unit 01:Completed 2200000 out of 10000000 steps (22%)
12:42:18:Unit 01:Completed 2300000 out of 10000000 steps (23%)
12:44:34:Unit 01:Completed 2400000 out of 10000000 steps (24%)
12:46:59:Unit 01:Completed 2500000 out of 10000000 steps (25%)
12:49:15:Unit 01:Completed 2600000 out of 10000000 steps (26%)
12:51:31:Unit 01:Completed 2700000 out of 10000000 steps (27%)
12:53:47:Unit 01:Completed 2800000 out of 10000000 steps (28%)
12:56:03:Unit 01:Completed 2900000 out of 10000000 steps (29%)
12:58:18:Unit 01:Completed 3000000 out of 10000000 steps (30%)
13:00:34:Unit 01:Completed 3100000 out of 10000000 steps (31%)
13:02:50:Unit 01:Completed 3200000 out of 10000000 steps (32%)
13:05:06:Unit 01:Completed 3300000 out of 10000000 steps (33%)
13:07:21:Unit 01:Completed 3400000 out of 10000000 steps (34%)
13:09:37:Unit 01:Completed 3500000 out of 10000000 steps (35%)
13:11:53:Unit 01:Completed 3600000 out of 10000000 steps (36%)
13:14:09:Unit 01:Completed 3700000 out of 10000000 steps (37%)
13:16:24:Unit 01:Completed 3800000 out of 10000000 steps (38%)
13:18:41:Unit 01:Completed 3900000 out of 10000000 steps (39%)
13:20:56:Unit 01:Completed 4000000 out of 10000000 steps (40%)
13:23:12:Unit 01:Completed 4100000 out of 10000000 steps (41%)
13:25:28:Unit 01:Completed 4200000 out of 10000000 steps (42%)
13:27:44:Unit 01:Completed 4300000 out of 10000000 steps (43%)
13:30:00:Unit 01:Completed 4400000 out of 10000000 steps (44%)
13:32:16:Unit 01:Completed 4500000 out of 10000000 steps (45%)
13:34:31:Unit 01:Completed 4600000 out of 10000000 steps (46%)
13:36:47:Unit 01:Completed 4700000 out of 10000000 steps (47%)
13:39:03:Unit 01:Completed 4800000 out of 10000000 steps (48%)
13:41:19:Unit 01:Completed 4900000 out of 10000000 steps (49%)
13:43:35:Unit 01:Completed 5000000 out of 10000000 steps (50%)
13:45:51:Unit 01:Completed 5100000 out of 10000000 steps (51%)
13:48:07:Unit 01:Completed 5200000 out of 10000000 steps (52%)
13:50:23:Unit 01:Completed 5300000 out of 10000000 steps (53%)
13:52:39:Unit 01:Completed 5400000 out of 10000000 steps (54%)
13:54:55:Unit 01:Completed 5500000 out of 10000000 steps (55%)
13:57:11:Unit 01:Completed 5600000 out of 10000000 steps (56%)
13:59:26:Unit 01:Completed 5700000 out of 10000000 steps (57%)
14:01:42:Unit 01:Completed 5800000 out of 10000000 steps (58%)
14:03:58:Unit 01:Completed 5900000 out of 10000000 steps (59%)
14:06:14:Unit 01:Completed 6000000 out of 10000000 steps (60%)
14:08:30:Unit 01:Completed 6100000 out of 10000000 steps (61%)
14:10:46:Unit 01:Completed 6200000 out of 10000000 steps (62%)
14:13:02:Unit 01:Completed 6300000 out of 10000000 steps (63%)
14:15:18:Unit 01:Completed 6400000 out of 10000000 steps (64%)
14:17:34:Unit 01:Completed 6500000 out of 10000000 steps (65%)
14:19:50:Unit 01:Completed 6600000 out of 10000000 steps (66%)
14:22:06:Unit 01:Completed 6700000 out of 10000000 steps (67%)
14:24:22:Unit 01:Completed 6800000 out of 10000000 steps (68%)
14:26:38:Unit 01:Completed 6900000 out of 10000000 steps (69%)
14:28:54:Unit 01:Completed 7000000 out of 10000000 steps (70%)
14:31:10:Unit 01:Completed 7100000 out of 10000000 steps (71%)
14:33:26:Unit 01:Completed 7200000 out of 10000000 steps (72%)
14:35:53:Unit 01:Completed 7300000 out of 10000000 steps (73%)
14:38:09:Unit 01:Completed 7400000 out of 10000000 steps (74%)
14:40:25:Unit 01:Completed 7500000 out of 10000000 steps (75%)
14:42:41:Unit 01:Completed 7600000 out of 10000000 steps (76%)
14:44:58:Unit 01:Completed 7700000 out of 10000000 steps (77%)
14:47:14:Unit 01:Completed 7800000 out of 10000000 steps (78%)
14:49:41:Unit 01:Completed 7900000 out of 10000000 steps (79%)
14:51:57:Unit 01:Completed 8000000 out of 10000000 steps (80%)
14:54:13:Unit 01:Completed 8100000 out of 10000000 steps (81%)
14:56:29:Unit 01:Completed 8200000 out of 10000000 steps (82%)
14:58:45:Unit 01:Completed 8300000 out of 10000000 steps (83%)
15:01:02:Unit 01:Completed 8400000 out of 10000000 steps (84%)
15:03:18:Unit 01:Completed 8500000 out of 10000000 steps (85%)
15:05:34:Unit 01:Completed 8600000 out of 10000000 steps (86%)
15:07:50:Unit 01:Completed 8700000 out of 10000000 steps (87%)
15:10:07:Unit 01:Completed 8800000 out of 10000000 steps (88%)
15:12:23:Unit 01:Completed 8900000 out of 10000000 steps (89%)
15:14:39:Unit 01:Completed 9000000 out of 10000000 steps (90%)
15:16:55:Unit 01:Completed 9100000 out of 10000000 steps (91%)
15:19:12:Unit 01:Completed 9200000 out of 10000000 steps (92%)
15:21:28:Unit 01:Completed 9300000 out of 10000000 steps (93%)
15:23:46:Unit 01:Completed 9400000 out of 10000000 steps (94%)
15:26:02:Unit 01:Completed 9500000 out of 10000000 steps (95%)
15:28:25:Unit 01:Completed 9600000 out of 10000000 steps (96%)
15:30:51:Unit 01:Completed 9700000 out of 10000000 steps (97%)
15:33:17:Unit 01:Completed 9800000 out of 10000000 steps (98%)
15:35:33:Unit 01:Completed 9900000 out of 10000000 steps (99%)
15:35:34:Connecting to assign3.stanford.edu:8080
15:35:34:News: Welcome to Folding@Home
15:35:34:Assigned to work server 171.64.65.104
15:35:34:Requesting new work unit for slot 00: RUNNING smp:8 from 171.64.65.104
15:35:34:Connecting to 171.64.65.104:8080
15:35:35:Slot 00: Downloading 331.28KiB
15:35:41:Slot 00: 92.64%
15:35:44:Slot 00: Download complete
15:35:44:Received Unit: id:00 state😀OWNLOAD error:OK project:7610 run:315 clone:0 gen:45 core:0xa4 unit:0x0000003e664f2dd04de6d3cc52b3848a
15:36:31:Slot 00 finishing
15:37:50:Unit 01:Completed 10000000 out of 10000000 steps (100%)
15:37:50:Unit 01😀ynamicWrapper: Finished Work Unit: sleep=10000
15:38:00:Unit 01:
15:38:00:Unit 01:Finished Work Unit:
15:38:00:Unit 01:- Reading up to 2153724 from "01/wudata_01.trr": Read 2153724
15:38:00:Unit 01:trr file hash check passed.
15:38:00:Unit 01:- Reading up to 223720 from "01/wudata_01.xtc": Read 223720
15:38:00:Unit 01:xtc file hash check passed.
15:38:00:Unit 01:edr file hash check passed.
15:38:00:Unit 01:logfile size: 78830
15:38:00:Unit 01:Leaving Run
15:38:01:Unit 01:- Writing 2480746 bytes of core data to disk...
15:38:01:Unit 01😀one: 2480234 -> 1768729 (compressed to 71.3 percent)
15:38:01:Unit 01: ... Done.
15:38:02:FahCore, running Unit 01, returned: FINISHED_UNIT (100 = 0x64)
15:38:02:Sending unit results: id:01 state:SEND error:OK project:7010 run:1 clone:157 gen:63 core:0xa4 unit:0x000000b40001329c4dfb9545537edbb8
15:38:02:Unit 01: Uploading 1.69MiB to 129.74.85.15
15:38:02:Connecting to 129.74.85.15:8080
15:38:08:Unit 01: 37.50%
15:38:14:Unit 01: 78.02%
15:38:22:Unit 01: Upload complete
15:38:22:Server responded WORK_ACK (400)
15:38:22:Final credit estimate, 4284.00 points
15:38:22:Cleaning up Unit 01


Ignore the smileys.
Click to expand...




It's so far pretty decent in folding, but when you want to compare to the core i7 2600k..... well.... it's not as good as people would hope.
 
Upvote 0 Downvote
as a really far-fetched theory, what would happen if samsung bought amd, samsung have some pretty sweet fabs already, i think theyre currently about to ramp 32nm, and from what ive seen they have a good history with new processes, and they have a ton of money to invest, imagine a samsung laptop where every single component was made by samsung, they could be far cheaper than everyone else and still make a profit
 
Upvote 0 Downvote
as a really far-fetched theory, what would happen if samsung bought amd, samsung have some pretty sweet fabs already, i think theyre currently about to ramp 32nm, and from what ive seen they have a good history with new processes, and they have a ton of money to invest, imagine a samsung laptop where every single component was made by samsung, they could be far cheaper than everyone else and still make a profit
Click to expand...

IIRC, the x86 licensing with AMD is non transferable, even when bought. So that would have to be some nasty negotiation between the buyer and Intel.

I'm pretty sure buying AMD is pretty hard 😛

Cheers!
 
Upvote 0 Downvote
http://vr-zone.com/articles/amd-starts-hiring-world-class-soc-engineers/14113.html

So they fire 10%, to "save" money and now they are hiring new people?

Thats even stranger. And plus "world class" SoC engineers.

Good luck AMD, whatever you are doing. To me it seems liek you are just going in circles.

They better hope that new CEO is worth something and not just pushing to a market that is a fad. Tablets are cool but I doubt they will be anything more than they are now. PCs, desktops to be specific, will always have a major place no matter what. That ans servers. I hope they don't lose focus on thier server products.....
 
Upvote 0 Downvote
Maybe they'll be better at branding memory in there name than counting transistors. 😛

Speaking of which, i've seamed to of crossed paths of them on newegg.

http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100007611%2050001028%20600006050&IsNodeId=1
Click to expand...


still , priced competitively, how about these sticks performance, do you have any link
 
Upvote 0 Downvote
still , priced competitively, how about these sticks performance, do you have any link
Click to expand...

Well beyond listing there specs, No. There no review of them but truth is, I doubt there will be much as most ram run similar to each other under there rated/list specs.

Now what i can say is at least we know who is making AMD ram just from AMD home page.

http://www.amd.com/US/PRODUCTS/TECHNOLOGIES/RADEON-MEMORY/Pages/system-memory.aspx#4

and those are Patriot

http://www.patriotmemory.com/products/groupdetailp.jsp?prodgroupid=225&prodline=9&group=AMD%20Memory%20-%20Entertainment%20Edition&catid=34

And VisionTek

http://www.visiontek.com/products/memory/amd-memory.html


Looks like AMD is wise on this one of having 2 manufactures to brand ram in there name compared to the cpu side of things. 😉



 
Upvote 0 Downvote
which was pertty much the point I was trying to make. Without competition, Intel has no desire or reason to be faster, instead of putting money into R&D, just sit back and collect money, raise prices, and rake in the profits.

Intel still appears to be removing features like hardware virtualization on its K-models for whatever reason, as well as support for it trusted execution technology. It's funny how we end up paying more for less when it comes to Intel's high-end, overclockable CPUs

Read more: http://vr-zone.com/articles/even-more-ivy-bridge-details-leak/14082.html#ixzz1fEFOlv8l
Click to expand...

And 1/2 the power ... lol, 95W TDP to 77W TDP, not seeing that one other than Intel's overhype of 50% less power (maybe they meant at idle.)
Click to expand...


Intel does have competition from ARM. ARM now has 64bit capability (ARMv8) and will be in servers before the end of 2012. ARM is scalable to 64 cores per CPU. Windows 8 will run on ARM so they have to keep improving x86.
 
Upvote 0 Downvote
^^ Or, they could ditch it, figuring that now that Windows is supporting other architectures, and the fact that Win8 is going to be the last "legacy" Windows OS [last 32-bit build], now would be a good time to debut a replacement architecture, and move on to something more modern. Especially is MS totally redesigns windows after Win8, which I suspect they will.
 
Upvote 0 Downvote
^^ Or, they could ditch it, figuring that now that Windows is supporting other architectures, and the fact that Win8 is going to be the last "legacy" Windows OS [last 32-bit build], now would be a good time to debut a replacement architecture, and move on to something more modern. Especially is MS totally redesigns windows after Win8, which I suspect they will.
Click to expand...


I'm sorry but unless they can run x86 on arm this will never ever happen. Lots of users hated Windows vista because their old software didn't work. Just think about it when none of the software/drivers today don't work on arm. This is why i don't see Project Denver doing well on windows 8 unless its on a tablet. This is one area i support Intel 100% they are right Arm is going to have issues with support for old software that many companies still use.
 
Upvote 0 Downvote
Actually I heard that one too. Right - IBM lost most of their PC marketshare during the '80s, to the likes of Compaq, HP et al. So they jettisoned CISC in favor of RISC, which is basically what Cell and PowerPC is, IIRC. But modern x86 is basically RISC-like too, once you get past the x86 legacy front-end. At least that's my understanding.

Anyway, considering that IBM sold off their PC division to Lenovo some 16 years ago, long before ATI was up for sale, it would be quite a jump for IBM to be interested in buying ATI back in 2005.
Click to expand...


There was never an official design designated as "CISC", merely "not-RISC". RISC isn't a design so much as it's a set of engineering concepts. First being to reduce the number of unique instructions to the bare minimum to save on CPU die space. Don't create a new specialized instruction if the task can be accomplished in a reasonably small number of existing instructions. Don't create multiple complex memory addressing modes, instead try to use a single linear method whenever possible. Do not have variable instruction execution times, all instructions should be executable within a single clock cycle. This make scheduling and prediction significantly easier then if instructions can take variable amounts of execution time. And finally make your instruction sets aware of multiple processing entities, basically have SMP capability built into the design.

Now these aren't hard and fast rules and few processor design's adhere to them perfectly (SUN SPARC is the most stringent) but their good guidelines for maximizing utilization of available die space. And since these design principles were made a pseudo-standard after the 8086 ISA was made, there is no was that Intel could of applied them. In actuality 8086 is a pretty decent language for accomplishing singular tasks, it was never envisioned to scale beyond a single main generic processor and a single specialized math coprocessor.

Modern x86 CPU design's are really a whole batch of specialized processors grouped together. Then another specialized processor that translates, predicts and schedules x86 instructions is used to feed the other processors using a vendor specific language that invariably follows the RISC principles. What we do now is translate x86 into Intel/AMD/Via specific instructions, execute them and return the results. Isn't the most efficient nor optimum method to go about doing it, but it works and maintains backwards compatibility. The consumer industry has been incredibly slow to accept having multiple ISA's for their software. JAVA was supposed to fix that and we see how that turned out.
 
Upvote 0 Downvote
I took an AM3+ 990X Chipset board recently, which cost me quite a buck, I wonder if AM4 offers something good to make AMD's system of future processor compatibility the Leader in marketting.
So to speak, I really hope Piledriver does something to perform awesomely, yet cost around a Phenom II's original cost radius to get people to purchase it.
 
Upvote 0 Downvote
since they publicly stated that they won't focus as much on desktop high end market, there is a possibility that piledriver (the one without igp like fx) could be delayed till q4 2012 or beyond. besides, they should release a new revision of zambezi first. then trinity, new zacate and bobcat apus.... getting to piledriver could take a while. then there is glofo's (worsening?) production capability. if there is a socket am4, and amd can consistently release it's prodcuts and everything goes according to amd's roadmap, it might not be available till q2-q3 2013.
 
Upvote 0 Downvote
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom