AMD's 128bit Bulldozer is a potential game changer

R

RussianHAXOR

Supreme [H]ardness
Joined
Mar 14, 2007
Messages
5,590
This article is fairly long and explains the current and past state of AMD's processors as well as talks about the future with APUs and Bulldozer. It is fairly long, but very detailed about the situation AMD is in and explains what needs to happen for them.

It is never boring in the world of CPUs. Regardless of who's on top, plans for next generations tend to excite everybody in the eco-system… if you deliver, that is. AMD had a lot of tough times of late, and lost a lot of good people due to lack of proper management. In this article, we bring you a look into the architecture that everybody in the industry has been impatiently waiting for. But this time, AMD cannot afford to fail.

The Ex-Alpha engineering teams lead by Dirk Meyer that created K7 and K8 architecture messed everything up with Barcelona/Agena and the infamous TLB-bug [Translation-Lookaside Buffer]. Shanghai/Deneb cleaned a lot of things up and AMD is back being competitive again, but Intel is pushing hard: Intel is operating in tick-tock architectural mode, and so far - AMD isn't able to answer back. K10 and K10.5 were nothing else but improvements over the K8 architecture. Last time we saw a completely new architecture from AMD, the stock market thought that an online dog-food shop was worth half a billion US$, mainstream media was touting that the world is going to end with that horrible Y2K bug... Yes, quite a long time ago. But before we dig into Bulldozer's architecture, let's set the record clear, with a simple architectural comparison between AMD and Intel.
Click to expand...

http://brightsideofnews.com/news/20...bulldozer-is-a-128-bit-crunching-monster.aspx
 
interesting read though amd needs to really look into their own version of hyper threading
 
Definitely interesting.

I don't think AMD needs hyperthreading so much as they just need a massive increase in instructions per clock, which is exactly what it looks like Bulldozer is aimed at. Hyperthreading would be nice, but it is not a necessity by any means.

I wonder if this will be AMD's next big step to counter Intel finally, much like beating Intel to the punch with 64-bit consumer-grade CPUs. If so, I'm wondering if they'll try for quad-channel DDR3, too...

No matter what, whatever it is they're cooking, it needs to be more exciting than Phenom II. Not saying Phenom II was bad, I'm saying it was pretty exciting but the next jump needs to be even better for them to really bring enthusiasts back to the AMD camp.
 
The website was created by theo valich and the article was written by theo valich. He has about as much credibility as CNBC does regarding the stock market.

I'd even go as far as to exclaim that it's all a lie.
 
ElMoIsEviL said:
The website was created by theo valich and the article was written by theo valich. He has about as much credibility as CNBC does regarding the stock market.
I'd even go as far as to exclaim that it's all a lie.
Click to expand...
Lol, thats quite a statement. Just curious what news network does have credibility regarding the stock market? :p ;)
 
maxius said:
interesting read though amd needs to really look into their own version of hyper threading
Click to expand...

Just what AMD needs right now, another reason for Intel to try to yank thier x86 extension away from AMD. Trying to mimick Hyper Threading right now would be corporate suicide for AMD. As it is Intel is breathing down thier necks about x86 useage rights and all that other happy hoo ha BS.
 
maxius said:
interesting read though amd needs to really look into their own version of hyper threading
Click to expand...


no thank you.. i do not want hyperthreading at all.. worthless excessive heat creating feature..

LAiN_OfTheWired said:
Definitely interesting.

I don't think AMD needs hyperthreading so much as they just need a massive increase in instructions per clock, which is exactly what it looks like Bulldozer is aimed at. Hyperthreading would be nice, but it is not a necessity by any means.

I wonder if this will be AMD's next big step to counter Intel finally, much like beating Intel to the punch with 64-bit consumer-grade CPUs. If so, I'm wondering if they'll try for quad-channel DDR3, too...

No matter what, whatever it is they're cooking, it needs to be more exciting than Phenom II. Not saying Phenom II was bad, I'm saying it was pretty exciting but the next jump needs to be even better for them to really bring enthusiasts back to the AMD camp.
Click to expand...

i remember reading some where that they were possibly going to introduce quad channel ddr3 with the Istanbul.. but im not sure if they scrapped the idea or not..
 
Hard and Confused said:
Just what AMD needs right now, another reason for Intel to try to yank thier x86 extension away from AMD. Trying to mimick Hyper Threading right now would be corporate suicide for AMD.
Click to expand...
Not really. Many other processors have SMT capabilities (like Power 5 and Niagra). The real problem is that AMD would need to overhaul its K1* architecture to support it, something it doesn't have the time or money to do.

Did anyone ask Theo what happened to that 30,000 point 3DMark06 score, reverse hyperthreading or any of the other trash he posts? :p
 
We could argue putting Nehalem is in the same basket as Core 2 architecture, since it contains numerous improvements and nicely copies DEC Alpha's design - yes, oh shocker. In case you don't know, the IMC [Integrated Memory Controller] was massively used in the Alpha micro-architecture, the fastest x86 architecture to date [it wasn't even x86, it ran a translation layer from RISC to x86 CISC], The Alpha architecture was sold to Intel and Dirk Meyer's Alpha team switched to AMD [and that's how Opteron came to life]. Then again, we almost got Intel's short-sighted vision of "NetBust will get us to 10 GHz". Luckily for our power bills, and the laws of physics made sure that 1kW TDP CPUs never came to market.

In case you're interested in Alpha 21264 and 21364, you might be interested to know that this decade-old CPU architecture featured a 10-channel RDRAM IMC, with two channels being used for redundancy. The remaining eight achieved higher bandwidth than Intel Core 2 Quad, a CPU released almost a eight years later. Now that we're done with this look into the past, it's time to take a good look at AMD's future.
Click to expand...

Theo is GREATLY mistaken.

I've explained this numerous times in countless forums. AMD's K7 & K8 and Hypertransport as a whole are directly derived from Alpha technology. K7 uses the EV6 bus from Alpha while K8 uses the EV7 bus (now known as Hypertransport).

It is AMD who are using Alpha technology in a nearly identical fashion. EV7 uses the SAME 3HOP design as Hypertransport. (While QPi is made up of two paths per link on a 2HOP design which is DRASTICALLY different).

DEC Alpha had licensed it's EV6 and EV7 technology to AMD. The Alpha architecture was sold, along with most parts of DEC, to Compaq in 1998. A few years went by and the Alpha technology was eventually bought out by Intel. Intel was then forced to allow licenses to AMD, Samsung and I believe IBM as well due to an SEC ruling.

Theo Valich is a fan of AMDs and always has been one of the worst kind and now he's allowed to spread his misinformation through a website while pretending to remain credible. It's also worth mentioning that Barcelona and newer AMD designs already use a 128bit FPU. He's hyping the FPU thinking it's the registers that are going 128bit. Classic blinded fan he is :)

http://www.techlawjournal.com/atr/80427intc.htm
http://www.kirps.com/web/main/_blog...lled-the-fastest-processor-in-the-world.shtml
http://en.wikipedia.org/wiki/DEC_Alpha
http://en.wikipedia.org/wiki/Digital_Equipment_Corporation

Very important that folks who want to know the truth visit: http://www.phys.uu.nl/~steen/web02/opteron.html & http://www.arcade-eu.org/overview/2005/opteron.html#opteron
 
Last edited:
sirmonkey1985 said:
no thank you.. i do not want hyperthreading at all.. worthless excessive heat creating feature..



i remember reading some where that they were possibly going to introduce quad channel ddr3 with the Istanbul.. but im not sure if they scrapped the idea or not..
Click to expand...

Only for the rumored 12 core, dual die version.
 
meh, AMD doesnt even need to take the performance crown with bulldozer... CPUs are plenty fast for 99% of users out there...

they just need to pump out something analogous to the HD4xxx series on the CPU side: offer 85% of the performance of their competitor, in a much simpler, smaller, cheaper to make, and easier to manufacture architecture....
 
Good for AMD, they need to bring out a processor that can actually compete with Intel's high end segment.
 
i can see how amd is going to compete in the future if they do not come up with a answer hyper threading. if you can get a 48 core/thread 4 socket server for the price of one i7 16 core /32 thread 4 socket server what would get... that answer is simple more cores for the buck. i have a funny feeling though after looking at amd's quarterly presentation in 2012 amd will have a new server microarchitecture that i hope will have a more than 1 thread per-core

http://www.betanews.com/article/The-plan-to-get-AMD-Opteron-back-in-sync/1240525554
 
Last edited:
LAiN_OfTheWired said:
Definitely interesting.

I don't think AMD needs hyperthreading so much as they just need a massive increase in instructions per clock, which is exactly what it looks like Bulldozer is aimed at. Hyperthreading would be nice, but it is not a necessity by any means.

I wonder if this will be AMD's next big step to counter Intel finally, much like beating Intel to the punch with 64-bit consumer-grade CPUs. If so, I'm wondering if they'll try for quad-channel DDR3, too...

No matter what, whatever it is they're cooking, it needs to be more exciting than Phenom II. Not saying Phenom II was bad, I'm saying it was pretty exciting but the next jump needs to be even better for them to really bring enthusiasts back to the AMD camp.
Click to expand...

Hyperthreading is a way of increasing the IPC - all it does is utilize pipe stalls for doing something useful rather than burn them. Essentially, HT is always going to provide a positive gain in performance, so it is something that would be beneficial for them to look at. There are a lot of situations like cache misses that result in stalls, and no matter how much you improve the pipe, you can't fix them. So to recover those cycles, you have to do something with them that doesn't depend on whatever's causing the stall - aka HT.
 
You must log in or register to reply here.
Back
Top