CPU is dead? APU is king?

Peter G. 11 Feb, 2008 at 7:15pm ET Article published in Tech

Comments

1. 
   11 Feb 2008 ~ 7:31pm Thrax Hype hype hype, we're not a failing company, hype hype, slideshow hype, slides, hype hype AMD.
2. 
   11 Feb 2008 ~ 7:41pm Thelemech

   Thrax wrote:

   Hype hype hype, we're not a failing company, hype hype, slideshow hype, slides, hype hype AMD.

   
   
   
   
   ....poor AMD! And you use to have their brand logo in one of your old Avatars, did you not Thrax? My how times have changed for AMDvsINTEL.
3. 
   11 Feb 2008 ~ 7:42pm Thrax Times change indeed.
4. 
   11 Feb 2008 ~ 8:12pm Winfrey who's apu?
5. 
   11 Feb 2008 ~ 8:39pm Leonardo Apu's Hardware......is where I got my DIY computer beginnings. :D:D
   
   
   Agreed with the above - it's AMD blather directed at Wall Street: We are relevant. We have a bright future. We are innovative. It is wise to purchase our stock. (gosh, I hope all that will be true again someday)
   
   The "APU." Hmm, if it's multi-core, will all cores work?
6. 
   13 Feb 2008 ~ 11:45am RADA

   Thelemech wrote:

   
   
   ....poor AMD! And you use to have their brand logo in one of your old Avatars, did you not Thrax? My how times have changed for AMDvsINTEL.

   
   I think that AMD has earned Thrax's ire over the constant posturing, flag waving, and the 3-CPU Monte game the company has been trying to dupe us consumers with since the arrival of Intel's Duo Core chip.
   
   Instead of AMD coming out and saying "Yup, Duo Core faster, we're behind and have made a few mistakes, but we're GOING to come back" its been 2+ years of " Pay no attention to the man behind the curtain!!!"
   
   I had every intention of building an AMD box for my next build, but that was because AMD 64 procs were trouncing anything Intel put out. Then Duo Core showed up. I have never been a Fanboi for either company, I am a consumer that wants the best product for my hard earned $$, so for right now, and for the foreseeable future I will have to go with an Intel proc.
7. 
   13 Feb 2008 ~ 12:39pm UPSLynx

   Thrax wrote:

   Hype hype hype, we're not a failing company, hype hype, slideshow hype, slides, hype hype AMD.

   
   You forgot the part about not knowing how to distribute their money in a shrewd manner.
   
   It's sad times are the way they are. ATI has been shambles, they've only just released a high end video card that can compete with the 8800 GTX - which is a friggin year old. Nvidia doesn't have the competition they had years ago, so we don't see nearly as fast a progression in hardware manufacturing.
8. 
   13 Jan 2011 ~ 3:28pm UPSLynx I find this thread ironic for numerous reasons, and extremely hilarious.
9. 
   13 Jan 2011 ~ 3:41pm Thrax THANKS DIRK
10. 
    13 Jan 2011 ~ 4:11pm AlexDeGruven This thread is now phenomenal due to its ultra-meta nature.
11. 
    13 Jan 2011 ~ 4:34pm BuddyJ Good game everyone.
12. 
    13 Jan 2011 ~ 5:14pm primesuspect Man, if only we still had the name APUsHardware... we'd be double relevant all over again?
13. 
    13 Jan 2011 ~ 5:49pm pigflipper

    primesuspect wrote:

    Man, if only we still had the name APUsHardware... we'd be double relevant all over again?

    
    Is it available?
14. 
    13 Jan 2011 ~ 5:55pm fatcat yes, relevant
    
    
15. 
    18 Mar 2011 ~ 11:12am ins4n17y

    UPSLynx wrote:

    You forgot the part about not knowing how to distribute their money in a shrewd manner.
    
    It's sad times are the way they are. ATI has been shambles, they've only just released a high end video card that can compete with the 8800 GTX - which is a friggin year old. Nvidia doesn't have the competition they had years ago, so we don't see nearly as fast a progression in hardware manufacturing.

    
    the second quote i'll deal with first. first off the radeon is highly competitive. remember the radeon 9700 series (aka codename khan aka r300 core) the main reason why at the time of this poster in 2008 claimed ati was uncompetitive was because ati was merging with artx another graphics firm. dave orton, the ceo of artx became the ceo of ati to come. with the mixing still on-going i can imagine the bump in ati product lines affecting its competitive ability with nv green guys in santa clara. second when they did have their stuff together, the first mergered design chip from artx+ati (array technologies incorporated) was the r300. may i remind you it was the most lived gpu of its time, with games 'playable' (not to you enthusiasts (playable to the general mainstream oem market) 7 years after the release of the die) since then ati has been competitive especially with its all in wonder series (which i am or was a collector of). since then amd has in many cases remained at the top of the performance high end of the market.
    
    radeon 9700, radeon x1800, (radeon 3000 series <----sucked for the reasons stated above except it was amd merging with ati rather than ati with artx) but since then:
    
    radeon hd 4000 series (destroyed nv)
    radeon hd 5000 series (destroyed nv) (ok gf 100 beat it by a slight margin and then)
    radeon hd 6000 series (aka 6990) <----just destroyed gf 100. fine gtx 590's gonna come out.
    
    just get a crossfire setup with 2x6990 and a 1200 w psu. then oc the **** out of them!!!!
    
    btw southern islands is going to release soon. (radeon 7000 series)
    
    nvidia is just patching up the gf 100 which was the 400 series that now became the 500 series. everyone in the industry knows nvidia has a history of renaming SAME architectures. ati on the other hand, i think they have, but not really.
    
    (for all of those nv fan bois i'll give them credit over the 5800/5900 series geforce fx series. YES they were faster than ati but at the cost of your hearing, nv basically pumped the voltages out of the die requiring a massively LOUD dustuster cooling solution.)
    
    ok now time for the real article:
    
    first off when posts like this come along aka thrax (Hype hype hype, we're not a failing company, hype hype, slideshow hype, slides, hype hype AMD.) and i quote, you have to sit back and think about it for a minute. these are the exact same people/engineers who did the impossible. even when intel tried hard to do 64 bit on 32 and claimed it was impossible (even the israelies failed at it), amd came out with amd64 architecture in 2003 and everyone knows how that went down. waaay waaay before any apu came out, i knew people from ati (because if you look at my profile i'm 30 min away from markham ontario (aka ati hq [which is now amd])) and i knew that they would make something amazing. amd isn't and has never been about hype, for instance when have you ever heard amd spend advertising dollars on markting campaign? (yeah thats what i thought, ok to be fair oh well thats 'causes amd's total net worth (roughly 7 billion us $ (hey i'm being generous)) is not even 1% (figureatively speaking (again with the metaphors) of intels throw away budget) (intel's net worth is approx 122 billion us$ again look on yahoo finance) - fine its not exact but its disproportionally HUGE compared to amds assets. the bottom line is amd can't hype. they are in no position and never have been in a position to do so. anything they tell you is real. if not, poof by their assets compared to their competitors like intel they're gone, just like that. secondly with a disproportional assets of that magnitude, i don't think amd would even spend profusely on any hype or advertising campaign. intel inside, mhz = faster? all these come from intel. amd is smart in spending what little they have to ACTUAL engineering. here, i'll explain it to you in excruciating detail.
    
    when amd says they are going to do something they are going to do it. no hype, just engineering. if you don't believe me on this just look at their fabs. in this case i'm going to highlight one example fab 30. anyone in semiconductors knows two things, before dirk came along amd sucked. yeah i've been there i owned a k3. (actually its really a k6-3) when dirk came along he changed everything. how was a company called amd run by sanders-iii at the time if any of you remember then by hector ruiz and then dirk. (but dirk was under ruiz at the timeframe i'm speaking of) the first thing i'm going to mention was dirks leadership. back in the k5/k6 or k'x' days or so on amd never had a history. everyone thought amd was always going to be the underdog. when the athlon released its architecture remained UNCHANGED for nearly a decade making socket a (socket 462) the most lived socket in mainstream computing history before the advent of amd64 in '03. the first point i'm going to make is that an unchanged architecture for nearly a decade went on to be the most powerful chip in history.
    
    the second thing i'm going to mention is fabs. back then with no history amd had one fab. fab 30. if you don't know already, amd supplied nearly 30% and more of the world's cpu marketed semiconductor dies with one FAB the fab 30. that's one hell of a yield, one hell of a efficiency and sources indicate that fab 30 was actually producing at 150% capacity and all spaces were 'utilized'. the point here is that amds fabs or fab i should say is 150% efficient and thus have a strong backbone (unlike nvidia whose first yields of fermi were less than 2%) strong backbone for manufacturing. the bottom line, whatever amd/ati (daamit) now designs they surely will have no problems making it manufacturable and mass marketable. whatever they conjure, they CAN and WILL manufacture it. if 150% efficiency isn't a excruciatingly painful testimony of amds engineering talent, i don't know i give up because i don't know what is. lastly i can tell you even now though amd is fabless, their yields still lead the industry because if not, they won't survive.
    
    now we have the apu. do you really think amds engineers who have survived this long being the underdog that they are would hype about an apu?
    
    my thoughts, its gonna kill. wth llano at 400 dx 11 full on hvd 3 shaders PLUS multiple x86 core with a internal crossbar switch and a redesigned powerplane at soi-nm global factories production....and thats just the start of it. llano was just the apetizer for the real thing.
    
    the full on zembezi and orochi will arrive next year. thats my prediction and i stake my reputation on it that most likely if not now the architecture that amd proposes with its apu will be the future of high performance computing.
    
    put it this way, compare say fastest ram say ddr4 say its at pc whatever 20 000+. compare that to l1 cache speed. theres no comparison. going from the cpu die's execution units onto off die thru the bus through the pcie controller to the videocard no matter how powerful will always introduce a performance penalty due to latency.
    
    however replace that l1 cache with hmm a powerful 400 shader videocard.....right in its place. now thats what i'm talking about. amd claims that intel though they are large and in charge have no clue how to make a crossbar switch that can handle gpu and cpu traffic at the same time. (so intel opted for a large l3 i think go google it)
    
    anyways just think about that. oh yeah amds hype hype hype just 'cause core came out'...oh i don't think so :P they were serious about creating a new athlon 64.
    
    alright guys back to 2003 when all the doubters came and everyone laughed at 64 on 32. now lets all do the same thing with the apu.......anyways lets just see what happens ok? as i said before can amd make a dream chip, yeah they can. (previous examples athlon, athlon 64) <--- these are the only two architectures amd has provided us regardless of codenames (what i'm saying is their stars cores aka k10's and onwards are 'relatively unchanged'. go look at a phenom die and compare it to an 2003 a64 die if you don't believe me). can they manufacture it? you betcha. why has amd stayed in second place? they did it on purpose. they were busy staying in second place knowing something big was in the works.
    
    quick recap:
    
    why did nv titanium kill ati? they were working on r300. (which became the longest playble gpu) (merger with artx + ati still ongoing)
    
    why did r600 fail? they were merging with ati like above. and then hd 4000 and so on. since then ati has never been defeated. (ok for maybe a few weeks/months on the quick release for a new generation) at the time of writing this article if 590 gtx beats 6990, i assure you radeon 7000 southern islands isn't far behind. (i say 4 months tops and i mean tops!!!)
    
    why did intel destroy amd with p1, p2, p3? dirk came in and cleaned up house. and in the end it was amd NOT intel who passed the 1ghz barrier with athlon (thunderbird).
    
    why did intel beat athlon (which was the longest socket (462) ever lived in mainstream oem computing) ? so amd could release a64.
    
    AND NOW why did intel beat a64? for the APU <---yeah well every compant leapfrogs. thats the point when amd/ati leapfrogs it usually takes 2-3 generations of nearly a decade for the competition to catch up. for instance the integrated memory controller in 2003 by amd, intel took until first half of 2010 to release i7 nehalem with intels FIRST imc.
    
    yes companies leapfrog but when amd/ati do it magic happens. its just fun to watch. now ati and amd both very talented and capable firms are now merged and people have doubts about the apu? get real. i have given you in this article their histories. i mean 30%+ supply of the world cpu market on one fab (<---- may i mention ONE fab no thats ONE fab like UNO ONE like yeah 30%+) who does that? intel has 6-12 fabs by comparison of that timeline (some of which are no longer operational (i knew someone at a fab that got closed down))
    
    so look deeply into their histories for their track record. and amd/ati did this at one point when i had nothing to write about. if they can do it once they will do it again. apu is king!!! finally once and for all i have settled this post.
    
    cheers
16. 
    18 Mar 2011 ~ 12:57pm Thrax what.
17. 
    18 Mar 2011 ~ 1:09pm ardichoke Way2respond to a 3 year old analysis.
18. 
    18 Mar 2011 ~ 1:12pm Basil

    ins4n17y wrote:

    wordswordswordswordswordswordswordswordswordswordswordswords

    
    
19. 
    18 Mar 2011 ~ 1:28pm PirateNinja

    Basil wrote:

    

    
    lols :smiles:
20. 
    18 Mar 2011 ~ 1:59pm Butters

    ins4n17y wrote:

    however replace that l1 cache with hmm a powerful 400 shader videocard.....right in its place. now thats what i'm talking about. amd claims that intel though they are large and in charge have no clue how to make a crossbar switch that can handle gpu and cpu traffic at the same time. (so intel opted for a large l3 i think go google it)

    
    I second this.
21. 
    19 Mar 2011 ~ 2:48am ins4n17y

    ardichoke wrote:

    Way2respond to a 3 year old analysis.

    
    actually to put it this way the apu debate is still an on-going thing, it may have started three or so years ago but the apu remains still for the most part unreleased. even still because of that speculation still remains and people are still unsure about the apu architecture.
    
    in point, to put it another way, all i simply did was shed some light between the histories of amd and ati at an attempt to "predict" (i use that loosely) that an one die gpu is the future for x86.
    
    so even though we are near to launch and the architecture still hasn't made its debut i am conclusively saying that amd and ati ill pull it off.
    
    thats my prediction anyways.
22. 
    25 Mar 2011 ~ 12:47pm GHoosdum

    primesuspect wrote:

    Man, if only we still had the name APUsHardware... we'd be double relevant all over again?

    
    I think the site will be changing names back to Apu's Hardware on April 1st this year, yes?
23. 
    25 Mar 2011 ~ 12:53pm primesuspect That would be too obvious. Plus, it wouldn't get enough stuffypants' to ragequit
24. 
    27 Mar 2011 ~ 4:59pm Garg How about only letting it render in IE6 on 4/1?
25. 
    28 Mar 2011 ~ 8:13pm AlexDeGruven Why would we only want to access the site?