AMD CPU speculation... and expert conjecture

[gamerk316]

And bringing up (again) the art of threading:

http://blogs.technet.com/b/winserverperformance/archive/2008/04/25/designing-applications-for-high-performance-part-1.aspx
http://blogs.technet.com/b/winserverperformance/archive/2008/05/21/designing-applications-for-high-performance-part-ii.aspx
http://blogs.technet.com/b/winserverperformance/archive/2008/06/26/designing-applications-for-high-performance-part-iii.aspx

I loved how my stance of "too many threads is bad for performance" is agreed upon by someone.

 

[cgner]

^ I agreed with that for the last 2 years lol

 

[griptwister]

oopsie... i guess amd doesn't pay the employees enough
"AMD accuses former employees of spying for Nvidia "
http://www.fudzilla.com/home/item/30140-amd-accuses-former-employees-of-spying-for-nvidia
glofo seems to be doing well recently.
http://www.fudzilla.com/home/item/30149-tsmc-still-leading-contract-chipmaker
http://www.xbitlabs.com/news/other/display/20130115235600_GlobalFoundries_Touts_10_Billion_Fab_8_Module_2.html

"It's just good business."

Lol, Pirates of the Caribean ftw! (purposely misspelled)

Like I said a while back, I have a friend that works over at Intel. He left AMD because they weren't paying enough... even a few years ago when they were doing good. They almost fired him for bringing an intel laptop into work (understandable though lol). Seems like the low payment of employees cost them employee loyalty. Sad situation though. AMD needs to step it up in the GPU department this year...

 

[noob2222]

he's lucky he didn't get fired for conflict of interest. Im pretty sure if you work for Intel, theyd fire you on the spot for bringing in an AMD powered product.

same basis if you work for coka-cola, you can't eat at pizza hut, a friend of mine got fired for that one.

 

[sonoran]

he's lucky he didn't get fired for conflict of interest. Im pretty sure if you work for Intel, theyd fire you on the spot for bringing in an AMD powered product.

Laugh about it, maybe. Fire you, no.

* Not speaking for Intel Corp *

 

[Cazalan]

he's lucky he didn't get fired for conflict of interest. Im pretty sure if you work for Intel, theyd fire you on the spot for bringing in an AMD powered product.

How else would Intel make sure their compilers work better for Intel chips if they don't have AMD kit to profile?

 

[JAYDEEJOHN]

Now, flip the marketshare

 

[blackkstar]

Worth a look: How a lowly E6400 holds up in gaming:

http://www.anandtech.com/show/6670/dragging-core2duo-into-2013-time-for-an-upgrade

Metro shows a clear GPU bottleneck, as a single 580 shows the E6400 is only 2 FPS slower compared to...an i7 3960X.

Granted, Dirt shows what you would expect, with the E6400 clearly lagging behind. But still...Very interested to see how a Q9550 is holding up now...

Uh, lets see. Released yesterday, missing FX 8350 in all benchmarks. FX 8150 missing in most multi-thread benchmarks, specially the ones it traditionally does very well in (x264).

But hey, don't forget! You're a complete idiot for thinking that reviews aren't fair! HAHA let's laugh at the idiots who think Anandtech has a thing against AMD!!!! They're so stupid!

It's like Anandtech just deleted FX 8150 from the graphs where FX did well. Can't have FX 8150 beating 2500k or 3570k!

This is AMD's biggest problem moving forward. The fact that the press will bend over for Intel and take it balls deep when they do things like announce a 7w CPU using SDP (which they just invented!) instead of TDP or when they show their new iGPU's gaming performance by playing a video in VLC and the guy on stage is pretending to play it.

You guys are all caught up in how the software is optimized, but AMD and Intel are different architectures. There's going to be software that just runs well on one and not the other and vice versa. The real problem is review sites that load their reviews so that they use benchmarks that favor a specific architecture.

I mean, if I took a graphics card review for Nvidia, and it consisted entirely of TWIMTBP titles, it'd be kind of shitty. Same with nothing but gaming evolved titles on an AMD graphics card.

Nobody holds sites responsible for their software choice. I mean, hell, most of Anandtech's bench is SYSmark. The same SYSmark they even wrote about becoming an only Intel piece of software: http://www.anandtech.com/show/4464/amd-resigns-from-bapco-over-sysmark12-concerns

The truth is that Nvidia, AMD, and Intel get to make optimizations to all these games, and then we look at the results of benchmarking these games and act like the software was created fairly. It's completely ******* idiotic.

 

[JAYDEEJOHN]

http://semiaccurate.com/2013/01/16/facebook-goes-arm-with-calxeda-like-semiaccurate-reported-2-years-ago/
Seems ARM could have some potential down the road for AMD

 

[jdwii]

he's lucky he didn't get fired for conflict of interest. Im pretty sure if you work for Intel, theyd fire you on the spot for bringing in an AMD powered product.

same basis if you work for coka-cola, you can't eat at pizza hut, a friend of mine got fired for that one.

I knew people who worked at GM and drove a ford everyday to work.

 

[jdwii]

blackkstar i have to say i agree with most of your statements. You can see how you are right at least about games a few pages back.(from a video)

 

[mayankleoboy1]

http://semiaccurate.com/2013/01/16/facebook-goes-arm-with-calxeda-like-semiaccurate-reported-2-years-ago/
Seems ARM could have some potential down the road for AMD

Since today is 17 Jan, the S|A articles before 17 dec have been "unlocked". You can read them.
Special mention : Broadwell chips will only have PCIE2.0-x4 lanes. So big dGPU's will be completely gimped.

 

[mayankleoboy1]

Meanwhile, AMD improves their "Minimum FPS measured in 1 frame interval" score :

http://techreport.com/discussion/24218/a-driver-update-to-reduce-radeon-frame-times#metal

Nice. If only the various dime-a-dozen review sites modify their methodology's...

 

[JAYDEEJOHN]

Meanwhile, AMD improves their "Minimum FPS measured in 1 frame interval" score :

http://techreport.com/discussion/24218/a-driver-update-to-reduce-radeon-frame-times#metal

Nice. If only the various dime-a-dozen review sites modify their methodology's...

The driver does not yet contain the new video memory manager. Our intention is release a new driver in a few weeks, which does include the new Video memory manager, which will help resolve latency issues for DX11/DX10 applications.
http://techreport.com/review/24218/a-driver-update-to-reduce-radeon-frame-times/5

 

[mayankleoboy1]

Quite late in the product lifetime cycle to release a new Video memory manager, no?
Or maybe GCN Refresh can benefit from that too...

 

[gamerk316]

I'm more worried that AMD didn't even bother to investigate these performance anomalies until TR brought the topic up. Heck, apparently even NVIDIA started only two years ago.

Still, glad to see this is FINALLY getting attention.



And in response to Blackstar: Doesn't the latest round of benchmarking by TR now put AMD ~ NVIDIA in these titles, minus 2-3 FPS or so? Maybe, I don't know, the DRIVER was the problem the entire time? If AMD has a driver problem (they did), I'd rather it be exposed and they fix it, rather then blaming NVIDIA and continuing to push a broken driver.

As a little FYI, NVIDIA and ATI/AMD have always had different approaches to GPU's. ATI/AMD always had the edge in memory bandwidth, NVIDIA always had the edge in shader performance. Problem for ATI/AMD, prior to DX9, performance was dominated by shaders, hence NVIDIAs dominance. But while NVIDIA was busy making the Xbox GPU, ATI/AMD successfully lobbied to get several features added to the DX9 spec whose performance was dominated by memory bandwidth. Low and behold, as a result, the NVIDIA FX series STANK in any DX9 benchmark, due to its poor memory bandwidth compared to AMD.

As a result, when you see similar performing GPU's (say, 7950 and 660 TI), and one easily outperforms on game "X", you can usually assume that game is driven by Shader performance (if NVIDIA is better) or memory bandwidth (if ATI/AMD is better). For the most part, there aren't that many tricks in the DX API you can use to move performance one way or the other. Aside from games that are dominated by one type of performance metric, you aren't going to see a lot of variance between similar GPU's of the same generation, regardless of vendor, unless theres an underlying driver issue.

 

[noob2222]

Quite late in the product lifetime cycle to release a new Video memory manager, no?
Or maybe GCN Refresh can benefit from that too...

HAHAHAHA and you try to call me out for not reading

Dodd says additional improvements are coming down the pike, including a rewrite of the software memory manager for GPUs based on the Graphics Core Next architecture that should bring a more general improvement:

 

[mayankleoboy1]

Wut ?
GCN = Graphics Core Next a.k.a HD79xx
GCN Refresh = HD89xx (to be declared)

Did you read what you linked to ?

 

[noob2222]

rofl, your an idiot. what does life cycle have anything to do with updating software drivers? or maybe you thought it was some sort of hardware change because you didn't read it yourself.

 

[mayankleoboy1]

what does life cycle have anything to do with updating software drivers?

sigh


1.

Quite late in the product lifetime cycle to release a new Video memory manager, no?
Or maybe GCN Refresh can benefit from that too...

2.

Dodd says additional improvements are coming down the pike, including a rewrite of the software memory manager for GPUs based on the Graphics Core Next architecture that should bring a more general improvement:

2+ (time GCN has been launched+time left till GCN refresh comes out) => 1

but you dont read....Worse, you purposely dont read.


Edit : let me slow it down and say in simpler terms : the driver update is specific to GCN based GPU's. Its been 1 year since GCN launched. And teh GCN+ arch is on the horizon. So it is surprising to see a rewrite of the VM manager, which by all means would be very very onerous task. They would possibly doing this only if the GCN+ can also possibly take advantage of the rewrite.
Also, HD6xxx and below wouldnt be advantaged by the rewrite.

Got it, Noob ? or you want me to rewrite this completely in words with 4 letters or less ?

 

[noob2222]

maybe you should try to explain wtf your smoking. From what I can see you beilieve that software updates will only affect old hardware? or that when new hardware comes out, that software updates should stop.

or mabye you think there is nothing at all in common between GCN and "GCN refresh" and that the software won't work at all.

or maybe your trying to justify your comment by arguement of insult.

your speculation that its pointless to update the software just because new hardware is coming out is borderline WTF.

edit: after your edit, got it, you believe that supporting older hardware is pointless.

 

[mayankleoboy1]

edit: after your edit, got it, you believe that supporting older hardware is pointless.

Lets have a poll on what I meant.
Hey noob, try this : http://www.amazon.com/Reading-Information-Elementary-School-Comprehension/dp/0131707493/ref=sr_1_1?ie=UTF8&qid=1358446188&sr=8-1&keywords=elementary+reading

 

[noob2222]

rofl,

I read an article that said something about GCN update, WTF why update something that old, it came out last year for crying out loud, support should totally be discontinued. I made up my own theory that GCN and GCN refresh can't at all work on the same driver set because there is nothing at all in in the article saying it can, its not even the same GCN core.

 

[mayankleoboy1]

cant parse gibberish.

Edit :

I made up my own theory that GCN and GCN refresh can't at all work on the same driver set

If you would have actually read what i wrote, you would have noticed i wrote just the opposite. This is what i wrote :

They would possibly doing this only if the GCN+ can also possibly take advantage of the rewrite.

But you dont read. And with every comment, you prove it further. You sure you arent retarded/6 year old with internet ? And please stop following what my signature quote says.

 

[noob2222]

cant parse gibberish.

Ok, ill quit trying to paraphrase your gibberish.

Quite late in the product lifetime cycle to release a new Video memory manager, no?
Or maybe GCN Refresh can benefit from that too...

They would possibly doing this only if the GCN+ can also possibly take advantage of the rewrite.

so now instead of maybe, its only if it can, wich proves my point, you think its pointless to update older hardware unless it affects new components, even if the new hardware isn't due for another 3ish months.

Makes it a little hard to update software for hardware that doesn't exist yet don't you think? But then again, your the one reading into the GCN update as being pointless since its late in the lifetime cycle, 3 months before the next hardware will even show up.