Still waiting on a replacement for my Phenom II x4 955 ... I love how they added more cores in the last 4 years, but the single threaded performance has barely grown to the benchmarks I could look up. Lets just hope their high dense library lib, and other enhancements will save speed-up their single core speed. Then they get more money.
Why am I waiting for AMD, and don't I buy an Intel? I'd like to support competition in the market.
If I'm misinformed and the new CPU's ARE faster, give me a poke.
The latest out of Tom's Hardware[1] seems to indicate that the combination of improvements in threading and modest improvements in the recent CPUs adds up to appreciable gains vs. the old Phenoms.
If you're laser-focused on single-threaded performance, steamroller is probably the one to look out for. They claim 30% improvement in ops per cycle.
3.2? My FX-8320 runs at 4.8ghz. Even if single threaded isn't improved, the headroom for a higher clock is there, thus giving overall better performance.
But you are probably pulling 160+ watts do accomplish that. On modern Intel parts, you can push 4.5 around 120 - 130. That means significantly less cooling requiremenets, so you could stick the chip in a smaller chassis, etc.
Ok, then I suggest you sift through more closely. More and more apps & games are leveraging multi-core (more than most people realize), so the charts in the link I included are an amalgamation of single-threaded and multi-threaded programs, which is how Tom's tries to put together their "real world" indexes.
But whereas Intel divested itself of its ARM-based XScale chip business several years back, AMD let it be known earlier this year that it planned to expand into ARM territory.
1) Bravo on moving to ARM, AMD!
2) It's not like you had a choice. :p
They're focusing on making themselves the premiere chip designer while Intel seems to be more strategically focused on manufacturing.
I would honestly like to see what happens if AMD makes a heterogenous CPU. one with say 4 arm cores and 4 x86 cores. where you could shut off the x86 cores if you didn't need them.
This could be usefull for High-Availability / Fault-Tolerant applications.
I think 10 years ago, Ericsson had a highly available server machine for Telecom application, named Jambala.
It had 2 SPARC and 2 x86 CPUs. Having two different ISAs in the same system can protect you from subtle software/hardware errors.
So you compile you code to both targets, if for some reason one of x86 CPUs fail, the second x86 takeover, if both x86 fail, then SPARC CPU takeover, etc.
One possible use case for heterogeneous CPUs that comes to mind is in laptops the same way AMD Switchable Graphics is used now: to save power when you don't need the performance.
However, I'm not sure if any energy efficiency benefits of this solution wouldn't be offset by the increased software and hardware complexity (which costs both money and battery power). If anything, it might be easier and cheaper for AMD to pair up its high-performance ("Steamroller") x86 cores with low-power ("Jaguar") ones instead.
Everyone's always had this idea, and it has very rarely proven to be terribly practical in reality. The few times that VMs of different hardware tend to be practical are when there is an enormous gap in performance between the emulated hardware and the host hardware (such as with old game consoles and arcade hardware).
I don't think it has been tried too many times in practice, because software binary translation is so proven (eg. like in android x86, or rosetta on mac, or fx!32 on nt, etc etc). Itanium is the only one that comes to mind. Any others?
Then there are several that have lower level "programmable microcode" kind of level in the hardware, ranging from the Transmeta stuff to mainframe-era writable control store...
That's not emulation that's decompilation / recompilation. Which is a very challenging task. TransMeta worked on this under the name of "code morphing". It's a promising technique but I don't think it has the maturity to be able to run cross-architecture code at close to hardware speeds (say, 50% or higher) without introducing new defects, especially x86 on ARM.
What you are discussing about, is a runtime issue.
I was discussing about installation time, which only prevents situations were you make use of code rewriting tricks, which are anyway frowned upon in the day and age of read-only code sections and no-execute data sections.
exactly. this is actually where it could end up really useful for windows 8 even. run unmodified rt versions of software or the original x86. enabling it could be very interesting for the surface tablet. it would also be in theory doable for osx and ios that way.
> The “Hierofalcon” series also provides enhanced security with support for ARM TrustZone technology and a dedicated cryptographic security co-processor
I'm always going to see stuff like that in a very suspicious way from now on, considering NSA has succeeded in putting backdoors in "encryption hardware".
> The “Adelaar” GPU family will deliver rich 3D graphics, multi-display support and support for DirectX 11.1, OpenGL 4.2
Sure, support the very latest DirectX, but only 2 versions behind OpenGL with your future new GPU. Way to go, AMD!
ARM Holdings is a French company. You probably don't need to worry about drop-in masks from them.
only 2 versions behind OpenGL...
It sure doesn't make sense why they didn't announce support OpenGL 4.4. I mean, the spec was released 6 weeks prior to the announcement- that's practically forever ago!
The older OpenGL version is probably just because of the driver, not the hardware. The latest Catalyst beta is about to bring 4.3 support (to almost all chip of the last ~3 generations). Just a matter of time for 4.4
Does anyone have any idea what the price on these things will be when they come out? I know AMD's x86 cpus are usually cheaper than Intel's, and given that ARM cpus cost less than x86 ones, this should also decrease the price.
recently really good progress has been made with the free amd drivers, also kernel 3.11 supports now the dpm (power management). also the uvd (video decoder) is working well. But if you want to play games or need maximum 3D performance there is no way around the proprietary fglrx driver from amd atm, although the free drivers keep getting better and better in this section.
Why am I waiting for AMD, and don't I buy an Intel? I'd like to support competition in the market.
If I'm misinformed and the new CPU's ARE faster, give me a poke.