Not too long ago these were their plans: http://www.cpu-world.com/news_2012/images/L_AMD_server_CPU_roadmap.jpg
"Improves performance/watt over existing 6300 family", much like Trinity to Richland. Not that hard to understand it, isnt it ??
Seems a bit hard for you to understand that I was talking about Centurion FX when I mentioned higher clocks at the expense of TDP. Also, if you didnt see any old chip, what would you call Warsaw? A new chip or an improved (better performance/watt) version of an old chip (like Richland and Trinity)?
Fits the the fact that there wont be any Steamroller-based Opterons with more than 2M/4C in 2014 (in other words: dont expect Steamroller based FX CPUs for desktop enthusiasts in 2014), so the best thay can do is OC the old chips at the expense of an insanely high TDP (meanwhile Intel releases 130W 12-core IB-EP CPUs for servers and a half harvested 6-core IB-E that unfortunately is enough to trash the competition performance wise for desktops).
All the points you make are so false, I'm just wondering if you are actually serious.
First, let's leave the marketing keyword ,up to, outside. In cold numbers reality, both Intel & AMD gained ~ 2-3% on Windows 8, on average, and that's what matters.
Second, Linux has not 8%. Head over to the Linux forums, there are threads discussing what Linux should do to go above 1% usage. You might want to go over there, and enlighten them (and the rest of the world) that they've been 800% wrong.
In the same time, I'm doubting you have used a single distro in your lifetime.
Third, CPUs are benchmarked with the most popular apps/games people use, and that's the way it should be. Therefore Windows & OS X get the priority here, noone cares what you personally want.
Fourth, I've built 4 hacintoshes, including my own, over the last 3 months, and AMD is a terrible no-go. Full of bugs, kernel hacks & works very unstable.
Quoting MacBreaker:
Mac OS X does not offer any official support for AMD processors, so the only methods available to AMD Hackintoshes remain in a state of quasi-legitimacy. For instance, the tonymacx86 forums expressly prohibit any discussion of AMD Hackintoshes, and even on forums that permit discussion, AMD is always a much less popular subject than Intel.
Keep in mind that support for AMD Hackintoshes is sparse, and in some cases nonexistent. Once you've installed Mac OS X on your AMD computer, whether with iBoot Legacy, nawcom's ModCD, or a distro, you're on your own.
Have fun sorting that mess out.
Fits the the fact that there wont be any Steamroller-based Opterons with more than 2M/4C in 2014 (in other words: dont expect Steamroller based FX CPUs for desktop enthusiasts in 2014), so the best thay can do is OC the old chips at the expense of an insanely high TDP (meanwhile Intel releases 130W 12-core IB-EP CPUs for servers and a half harvested 6-core IB-E that unfortunately is enough to trash the competition performance wise for desktops).
Your own words in response of a future FX line(2014) was an OC old Chip like the FX Centurion. Those Centurions are not the next desktop FX chips and since Warsaw is a new die its not an old chip no matter if its a new architecture or not, i dont suppose you call Richland an OC old chip.
Same die size as Trinity, same number of cores and SPs count. Yes, its an improved version of old Trinity, not exactly a new chip.
The term "average" can be a marketing keyword. Depending of the selected applications you can vary the average to one side or to the other side of a marketing goal.
I said "about" a 8%. Other estimations give 10%, for instance. The 1% myth has been debunked often.
You use again a priority argument. Debunked before in this same thread.
Apple does not authorize the use of the Mac OS on any x86 PC other than the ones it has developed itself. That is the origin for the term "hack" in hackintosh. As said before you can have an AMD hackintosh working, it is just needed a bit more of work because departs more from Apple hardware. Using this for you anti-AMD attack is kind of funny.
Apart from ignoring to comment on the February 2012 slide given, what is your point? That AMD is releasing Centurion FX line in 2013 because you claim there is no FX steamroller in 2014?
1) Average is what matters, and it has been proven many times that whatever 2-3%,if any, gains for windows 8 are both for Intel & AMD cpus, therefore using Win7 or 8, will not make or break the results. An i7 destroying the FX in Win7, will do the same in Win8 and by the same margin.
2) Your estimations are well known to be delusional. Linux has 1% market share, if you're in any doubts join any conversations in Linux forums.
3) Popularity > All. You arguments are beeing debunked constantly by other members here, and you continue to bury your head in the sand.
4) You said "Hacintosh runs fine on AMD cpus", another delusional statement, and I just stated why you were wrong. Apple does not use AMD cpus because of their terrible perf/watt. That's not an attack, that's a fact. You might wanna email them and enlighten them also.
1) Average is what matters, and it has been proven many times that whatever 2-3%,if any, gains for windows 8 are both for Intel & AMD cpus, therefore using Win7 or 8, will not make or break the results. An i7 destroying the FX in Win7, will do the same in Win8 and by the same margin.
2) Your estimations are well known to be delusional. Linux has 1% market share, if you're in any doubts join any conversations in Linux forums.
3) Popularity > All. You arguments are beeing debunked constantly by other members here, and you continue to bury your head in the sand.
4) You said "Hacintosh runs fine on AMD cpus", another delusional statement, and I just stated why you were wrong. Apple does not use AMD cpus because of their terrible perf/watt. That's not an attack, that's a fact. You might wanna email them and enlighten them also.
so you work at apple now? define this terrible perf/Watt.
ROFL.
I'd define terrible performance per-watt as a 125-watt CPU performing poorly compared to a 77-watt CPU.
I'd LAUGHINGLY define terrible performance per-watt as a 220-watt CPU breaking even with an 84-watt CPU.
Daimon
I think its safe to safe that plenty of apps on osx could have higher performance than its windows equivalent.
ROFL.
I'd define terrible performance per-watt as a 125-watt CPU performing poorly compared to a 77-watt CPU.
I'd LAUGHINGLY define terrible performance per-watt as a 220-watt CPU breaking even with an 84-watt CPU.
Daimon
Like it or not. Sweeper is right on this one. FX line seems to basically not carry over the Steamroller CPU cores like prior ones.
Same die size as Trinity, same number of cores and SPs count. Yes, its an improved version of old Trinity, not exactly a new chip. Same Piledriver cores unlike Haswell and Kaveri (new architecture).
The fact that there is no Steamroller-based Opteron with more than 2M/4C in 2014 kinda tells you that their priority is APUs. I dont believe they would make a new non-APU chip exclusively for high-end desktops and not release it for servers, but... believe what you want.
If they really wont have a Vishera successor in 2014 the best they can do is improve what they currently have (+clocks), unfortunately at the expense of an insanely high TDP (talking about Centurion vs ''regular'' FX here). Ps: I never said Centurion was their new FX line, I said it is the best they can release right now if Steamroller-based FX chips are not launching so soon... I'll believe you were not trolling but read carefully next time.
You cannot have a Richland APU like A10-6700 out of a Trinity chip, the two are not the same no matter if they have the same core and Cache count.
Since they will produce 4Module 8Core dies to use them for the server MCM 12-16core parts, they can easily make a 4Module 8Core desktop CPU using the same dies. Just because they will not use them in servers doesnt mean they will not have a desktop CPU.
Either you didn’t know they are using 4Module 8Core dies for the 12-16Core Server CPUs that make a warsaw 4Module 8Core desktop CPU possible or you specifically tried to pass the fud that they can only bring Centurion like CPUs in the Desktop for 2014.
I have given benchmarks where the 125W cpu is faster than the i7-3960X cpu (130W) and would be faster (my estimation) than the i7-3970X (150W).
Excellent post. Thus my request of galego to post his specs as you and I have.You compile whatever benchmarks you like, and run them on your 8350, say, at 4.8GHz. I'll run the standard binaries on a 3930K at stock. We'll post them here, with screenshots. I'll run them on my 3470 as well, if you like. I look forward to it. Without hardware at your fingertips, you're blowing smoke. I don't believe you have access to anything even approaching the hardware you espouse, and being on this forum all day, posting in a constant stream of free time, lets me understand why.
Daimon
Still waiting for a link with official confirmation of that there is no Steamroller FX line.
There is no steamroller FX line until AMD actually announces one, it doesn't exist. Can you prove otherwise with official confirmation?
AMD has stated that the FX-series will receive a Socket AM3+ version of a Steamroller based CPU in 2014
url]http://en.wikipedia.org/wiki/Bulldozer_(microarchitecture)#3rd_Generation_.28Steamroller.29[/url]
Could all that people who has negated FX steamroller provide some link please? If AMD has cancelled them some link would be available.
Did you bother check the reference links in the Wikipedia article you posted here? No? Because it remits you to FAD 2012 slides, which are far outdated by any standards, but, guess what, not even there Steamroller FX is mentioned.