Slashdot Mirror
AMD FX-8350 Review: Does Piledriver Fix Bulldozer's Flaws?
An anonymous reader writes "AMD just officially took the wraps off Vishera, its next generation of FX processors. Vishera is Piledriver-based like the recently-released Trinity APUs, and the successor to last year's Bulldozer CPU architecture. The octo-core flagship FX-8350 runs at 4.0 GHz and is listed for just $195. The 8350 is followed by the 3.5 GHz FX-8320 at $169. Hexa-core and quad-core parts are also launching, at $132 and $122, respectively. So how does Vishera stack up to Intel's lineup? The answer to that isn't so simple. The FX-8350 can't even beat Intel's previous-generation Core i5-2550K in single-threaded applications, yet it comes very close to matching the much more expensive ($330), current-gen Core i7-3770K in multi-threaded workloads. Vishera's weak point, however, is in power efficiency. On average, the FX-8350 uses about 50 W more than the i7-3770K. Intel aside, the Piledriver-based FX-8350 is a whole lot better than last year's Bulldozer-based FX-8150 which debuted at $235. While some of this has to do with performance improvements, that fact that AMD is asking $40 less this time around certainly doesn't hurt either. At under $200, AMD finally gives the enthusiast builder something to think about, albeit on the low-end."
Reviews are available at plenty of other hardware sites, too. Pick your favorite: PC Perspective, Tech Report, Extreme Tech, Hot Hardware, AnandTech, and [H]ard|OCP.
90+% of my CPU is idle time.
How much power does the new chip use at idle and how does that compare to Intel?
50W at the top end means about $25/yr if I was running it 24/7. But since typical desktop is idle, what is the power difference there??
And yes, I don't care about single thread performance as I care about multithread performance. Single thread performance has been good enough for desktop for almost a decade, and the only CPU intensive task I do is running those pesky `make -j X` commands. No, not emerging world or silly things like that ;)
I put together an 8way,32GB machine (no local storage) for $400 to play with ESXi. Courtesy of the freebie VMWare download and a reasonably priced 8way machine, I can get into some pretty serious VM work without spending a ton of dough. I don't need massive performance for a test lab.
I want to delete my account but Slashdot doesn't allow it.
These chips "excel" at big, heavily threaded workloads. Which is to say that they can beat similarly priced Intel chips that are simply up-clocked laptop parts. Move up to a hyperthreaded 3770K (still a laptop part) and Vishera usually loses. Overclock that 3770K to be on-par with the Vishera clocks while still using massively less power than Vishera and the 3770K wins practically every benchmark.
Unfortunately, if you *really* care about those workloads (as in money is on the line) then Intel has the LGA-2011 parts that are in a completely different universe than Vishera, including using less total power and being much much better at performance/watt to boot. I'm not even talking about the $1000 chips either, I'm talking about the sub $300 3820 that wins nearly every multi-threaded benchmark, not to mention that $500 3930K that wins every one by a wide margin.
So if you want to play games (which is what 90% of people on Slashdot really care about): Intel is price competitive with AMD and you'll have a lower-power system to boot. If you *really* care about heavily-multithreaded workloads: Intel is price competitive because the initial purchase price turns into a rounding error compared to the potential performance upside and long-term power savings you get with Intel.
Vishera is definitely better than Bulldozer, but AMD still has a long long way to go in this space.
AntiFA: An abbreviation for Anti First Amendment.
You do realize that most of your processing occurs via your gpu in a game right...?
New AMD processor, higher clocks than the last one but no massive improvements performance-wise. Still rocks at multi-threaded, integer-only workloads, still sucks at single-threaded or floating-point performance, still uses a huge amount of power. AMD giving up on the high end, their top-end parts are priced against the i5 series, not the i7. Since Intel's overpricing stuff, they're still roughly competitive. Might be good for server stuff, maybe office desktops if they can get the power down, but not looking good for gaming. Overall mood seems to be "AMD isn't dead yet, but they've given up on first place".
There. Now you don't need to read TFAs.
I've read through some of the reviews. It looks like a nice CPU with a bit too high power usage for my taste.
And please take benchmark results with a pinch of salt- most of them are compiled with Intel compiler, and will have lower results on AMD CPUs just because Intel compiler will disable a lot of optimizations on AMD CPUs.
I don't know of any site which would have Java application server, MySQL/PostgreSQL, python/perl/ruby, apache/PHP, GCC/llvm benchmarks under Linux. Video transcoding or gaming on Windows is really skewed and nowhere near to what I do with my machine.
--Coder
Here are a set of benchmarks that are more centered on the Linux world from phoronix and are thus a little less prone to intel compiler discrimination. The results seem more realistic: better and worse and similar to an i7 at different work, still hard on power usage, low purchase price.
"but money is the God of Algiers & Mahomet their prophet." - Rich. O'Bryen June 8th 1786
The new Trinity, and now these FX Procs, are perfect for "the 99%," that is to say for what 99% of people do with their machines: surf the web, check email, maybe do some photo editing or piecing together home movies.
They're cheap, reasonably fast, and support all the latest extensions and optimizations. Plus, even for enthusiast prosumers who want to screw around with things like virtualization, you can get into the IOMMU space cheaply with AMD, which is nice for a platform like ESXi that has a mature PCI passthrough implementation.
I think the Trinity A-10 does really hit the sweet spot for most consumers. It has reasonably fast processing and reasonably fast GPU in a ~$130 package. If you're putting a box together on the cheap, it becomes quite compelling.
I read phoronix a lot these days. I find more technical news in there than in Slashdot.
However, their benchmarks are often flawed. For example they did a Linux scheduler benchmark recently which measured throughput (average this or that) and not latency/interactivity (response times), which was totally useless. Well, ok, you can consider it a test checking for throughput regressions in interactivity oritiented schedulers, but it did not measure interactivity at all.
And regarding their Vishera benchmark, they measured most of their standard stuff, mostly scientific calculation, video/audio encoding, image processing, rendering. I very rarely do any of this.
The developer related benchmarks they had were Linux kernel compilation times (Vishera won), and you might count OpenSSL as well. They didn't do PostgreSQL, they didn't benchmark different programming languages, nor application servers, nor office applications, nor anything that would really interest me. I wish someone would measure Netbeans/Eclipse and other IDE performance.
And anyway, did you notice that AMD usually does much better in Phoronix reviews than in Anandtech/Toms Hardware/whatever sites? That's because Phoronix doesn't use Intel Compiler nor Windows, so results are much less skewed.
--Coder
CPU's tends not to be the bottle neck these days its either the hard drive or slow internet connection.
AMD fanboys have been saying that for years as AMD lagged behind Intel's CPU performance. But they then keep telling people they should buy AMD 8-core CPUs instead of Intel 4-core, when, if they really believed what they were saying, they would be telling people to buy the cheapest dual-core available, or an Atom.
Seriously, why would you possibly think an 8-core CPU was 'perfect' for your parents' PC, no matter who makes it? If they're not CPU-limited, buy the cheapest CPU you can get.
AMD FX-8350 Review: Does Piledriver Fix Bulldozer's Flaws?
No. It still guzzles power like crazy compared to Sandy/Ivy Bridge, and its single-threaded performance still sucks royally. (And that's still very important since many, many programs cannot and will not ever support full multithreading.)
For non-technical parents and other users, I actually like lots of cores and lots of memory-- more so than 'power users.'
Have you ever seen some people boot up their machines? It will take 5 minutes because of the sheer amount of crap installed.
It will be shit like two different anti-virus suites (the first one's subscription expired, and they installed a new one without uninstalling the old). There will be a update checker for every conceivable thing that's been installed. There's the preloader software that loads whatever crappy application it is into memory at boot time so that the application starts faster when you click on it's associated file type. Then, there are all the various helpful toolbars for Internet Explorer that "enhance your internet experience" and delivery dollars to your inbox. Of course, the user will probably go through two or more inkjet printer/scanner/copier all-in-one devices that require the manufacturer's 300 MB "printing experience suite" to enhance user experience by making pop-up windows when the ink is low and will helpfully tell you where you can order more special photo printing paper. There's the two different desktop search tools that got installed when the user downloaded the kitten screensavers pack and accompanying mouse cursors. Then, there is the sidebar app showing the weather outside and what time it is in Fiji.
That's stuff I can think of off the top of my head. I have no idea how average users manage to cruft up their machines so much, but they do. Keep in mind also that they're probably not going to upgrade hardware for a few years, but they'll likely install iTunes 32, Internet Explorer 17, and Office 2016, which will have >1 GB memory footprints.
By having lots of cores, and lots of memory, you can give them a decent user experience even on a machine that they've bogged to shit.
Have you ever seen some people boot up their machines? It will take 5 minutes because of the sheer amount of crap installed.
So uninstall the crap or install an SSD. A faster CPU makes far less difference to boot time than an SSD, because boot time is mostly limited by the seek performance of the disk as it loads all that crap into RAM.
Man I really want AMD to win!
I am typing this on a phenom II which is a better chip in my opinion and fast at the time in (unfortunately in 2010 standards). But these things run well over 130 watts, are loud with huge freaking fans, 4.4 ghz, and it seems AMD is trying to pump out as much speed as possible to beat intel's lowest end chips.
Just call it pentium IV 2.0 while we are at it? I am not a fan of intel because I run vmware and hate that intel cripples its chips and the bios to exclude virtualization on all but the most expensive units. I hate the cost of a high end icore 7 which in 2010 was only 10 - 15% faster than a Phenom II but cost 400% more where I can buy a whole system for the cost of a single intel core 7 extreme.
Well gentlemen. Expect dark days ahead and a return to $1000 desktops, $500 chips, and virtualization only available on xeon chips by next year. :-(
With AMD junk status it is bound to happen now since these chips can't match intels offering.
http://saveie6.com/
Actually check out the Anandtech review. They do it on windows 8 with the new scheduler... the difference in performance of the FX series from Windows 7 to Windows 8 is fucking mind-boggling. The FX-8350 actually trounces all but the higher end I5s and gives the I7 a run for its money in several games. For Sub-$200 if you're getting Windows 8(which I apparently will be now) you can't get anything thats even close performance wise.
I can't believe people keep on complaining about the bulldozer's FP performance. Does nobody realize the shared floating point unit is a 256-bit one, TWICE AS WIDE as the usual 128-bit? Also, assuming you're an HPC user, did you run actual benchmarks? Because we did, and for our (only modestly parallelizable) HPC workload, compiled with a bleeding-edge compiler (not Intel) that supports AVX and running on a bleeding-edge Linux kernel, Bulldozer was remarkably competitive with Intel's offerings at the time, with Interlagos and Westmere getting about the same amount of useful work done per clock cycle. There are some HPC benchmarks on AMDs website that seem very unlikely in light of the mainstream press. However, in light of our own benchmark results, they seem quite reasonable (although we never quite could make it look that good for AMD; probably because AMD didn't go to the same lengths to squeeze the maximum performance out of the Intel systems). Either way, AMD simply blew Intel away on a per-node-price basis, even when compared to Romley. All the way, I was the one arguing that "we should try Intels" based on reviews I saw online, but once we got all the benchmark results in, I simply couldn't argue anymore.
Also, if AMD's FP performance is truly that abysmal, please explain this? AMD bribed Dell more than Intel so that they now market Bulldozer-based Opterons as "excellent for oil and gas exploration, scientific and medical research, video rendering and other challenging HPC projects"???
Of course, this is all for a very specific workload and may not hold for all HPC workloads, but I have a strong feeling that even generally spoken, the Bulldozer's FP performance for HPC applications is just fine. It's just that most FP-intensive applications used in most of the benchmarks we're seeing in "end-user" space are not compiled to take full advantage of it and/or not running on an Operating System that takes full advantage of it and/or not very relevant test cases for the Bulldozer's parallel HPC potential. For example, one of the things we found out is that Intel's frequency scaling is more aggressive than AMD's, so Intel suffers badly if you put all the cores on a die to work at once. Also, Intel's improved HyperThreading still ain't worth shit if you saturate the FP units, while AMD's "clustered multithreading" succeeds to squeeze out a significant advantage owing to the fact that not all of our FP code is easily vectorizable so that sharing the 256-bit FP unit between 2 execution threads works better than trying to keep it busy with 1 thread's vectorized instructions.
/rambling rant
When you call an 8 core 4GHz CPU low-end.
How the heck is using a higher resolution relevant to a _CPU_ test? High resolutions stress out the GPU way more than the CPU.
You use a low resolution so that the bottleneck becomes the CPU and not the GPU.