Yes if you need more memory bandwidth, POWER is where its at. OTHO, as I've been saying the base cores don't do as well on single threaded loopy or problems with a decent L1/L2 cache hit rate. The specint_rate numbers look good because of the x8 threading, the single core results probably don't look that good.
So for many workloads it won't be that great. For the one I was working on out of the box, the power was 1/2 as fast. But that wasn't fair because we had a couple highly optimized x86 code paths. Doing some basic POWER optimizations brought the performance in line with the x86. But on some benchmarks it would win, and lose on others. So while we utilized a _LOT_ of memory/IO bandwidth the fact that there was nearly 2.5x available in the power system over our E5 didn't give us enough of a boost to make it worth the higher price tag (nearly 3x in our case, cause we were comparing with a supermicro machine). Maybe this newer power machine changes that a little.
Until about 5 years ago the Power architecture dominated in the CPU performance domain. Saying "IBM is too incompetetent" makes you sound like you don't really know this market.
Because I really dont. Yes Power used to dominated performance. I have always thought with every tick and tock Intel has closed the gap or leap forward already.
Intel is still winning / dominating, there were talks of Google using POWER8, haven't seen anything more. And knowing Intel with a clear Roadmap coming up.
Do Power8 have any place in Cloud Computing?
I too, wish Intel would have more competition, but i dont see the choice of using it. Most of the distributed and In Memory Processing aren't using it. But then I guess i am not the target audience.
In general, what's kept POWER where it is is that you need to be willing to pay far higher prices than Intel charge, and have far higher power consumption than high-end Intel parts, and have the cooling to cope with far higher exhaust temperatures.
Intel's only got the ability to develop one microarchitecture family. When they try for more than one, things fall through the cracks and you get the Pentium 4 dead-end or the Atom's lackluster performance.
Intel's primary microarchitecture is aimed at laptops. It scales reasonably down to large tablet power levels and up to workstation power levels. For the high-performance server market, they can only throw cores and cache at the problem, with some enterprise features bolted on.
IBM's always targeted the high-performance server market. For a while, their Power cores were also being used for workstations and even a desktop by Apple, but that's never been the focus. They include things like decimal arithmetic and SMT and hardware transactional memory and they've been selling the high-end parts at 4-5GHz speeds for a long time.
OK, what's Intel missing for high-performance servers? They've got respectable performance, they've got VT-x and all the other virtualization hardware, what's missing?
The x86 architecture was not amenable to a high performance pipelined implementation, so what Intel did since Pentium Pro is to JIT the x86 instructions into an internal RISCy instruction set that can be executed out-of-order with competitive performance.
The x86 architecture was limited to 32-bit, severely limiting the virtual address space as well as hampering OS implementation with hacks like PAE, so what AMD did is to extend it to 64-bit.
The x86 architecture was not designed for SMP scalability due to the rather strict memory ordering requirements, but most of the architectures with laxer memory models (in particular, Alpha, which was the most lax of all) are out of business today (in fact, of the commercially relevant server architectures today, only POWER has lax memory ordering; SPARC can in theory but is usually configured to run with TSO which is similar to x86).
The x86 architecture was not designed for OS virtualization because various instructions did not trap when executed in user mode, so what Intel and AMD did is define a new protection level ("ring -1") to run the hypervisor so this works efficiently now.
What actual problem do you see with x86 that cannot be solved or worked around by some creative engineering at Intel/AMD?
> The x86 architecture was not amenable to a high performance pipelined implementation, so what Intel did since Pentium Pro is to JIT the x86 instructions into an internal RISCy instruction set that can be executed out-of-order with competitive performance.
POWER does the exact same (I can't remember which revision).
It's probably worthwhile to point out that Itanium still, just about, exists, and has traditionally been Intel's competitor to POWER. Though certainly, in recent years, x86_64 has largely taken over that role.
Last I knew it was still close to a billion dollars of hardware per year—and even at the high prices Itanium is at, that's still a considerable number.
