The Vishera Review: AMD FX-8350, FX-8320, FX-6300 and FX-4300 Tested
by Anand Lal Shimpi on October 23, 2012 12:00 AM ESTProjected Performance: Can AMD Catch up with Intel?
I keep going back to this slide because it's incredibly relevant. It's the only indication we have from AMD of what its future roadmap will look like in terms of performance increases:
Each year AMD promised to increase performance of its high-end cores by roughly 10 - 15%. Astute observers will note that, at this rate, AMD will almost never catch up to Intel. AMD at the time was careful to point out that it's talking about 10 - 15% gains in core performance, and it could potentially see even larger increases in total chip performance by pulling other levers. Vishera is an example of AMD doing just that. The Piledriver cores by themselves don't increase performance tremendously, but they do give AMD a little more thermal headroom to work with thanks to some more efficient design decisions and better transistor choice. With Vishera, AMD took the additional power headroom and turned into a frequency advantage. The result is AMD's FX-8350 can operate in the same power envelope as the outgoing FX-8150, yet runs at an 11% higher base clock (turbo frequency remains the same). Through frequency and core level improvements, AMD was able to deliver a bit more than the 10 - 15% performance increased in promised.
If AMD is able to repeat these improvements again next year, I wondered whether or not it would get any closer to closing the gap with Intel - particularly when it came to single threaded performance. We already know from our Haswell investigations that Intel is expecting around a 5 - 15% increase in CPU performance from Haswell over Ivy Bridge. If we assume that Haswell delivers towards the 15% end of that spectrum, and if we assume that Steamroller delivers the same level of improvements that we saw from Piledriver/Vishera, we end up with some pretty interesting predictions for where things end up next year. I modeled the 2013 performance of high-end AMD and Intel platforms based on those two factors and plotted the curves in a few different benchmarks. For each generation I used the parts that AMD stacked up against one another (they are also fairly similarly priced). For 2011 I used the FX-8150 vs. Intel's Core i5 2500 and for 2012 I used the FX-8350 vs. Intel's Core i5 3570. The 2013 data is of course projected based on a 15% increase in performance from Haswell, and a repeat of the Vishera vs. Zambezi increase for AMD. This is mostly an interesting experiment so don't get too invested in the data.
We'll start with Cinebench, by far the most painful of the tests for AMD from a single-threaded performance perspective:
The Vishera gains here were decent but not enough to dramatically shrink the performance gap. Furthermore, Intel put a good amount of distance in place with Ivy Bridge and if it can continue that with Haswell I don't see much hope here.
The multithreaded Cinebench results begin in AMD's favor and remain so even with our projected performance data.
Mozilla's Kraken benchmark is another example of single threaded performance gone awry for AMD.
Thankfully, Vishera does close the gap by a decent amount and if AMD extends those gains it is on an intercept course with Intel. The bad news is, that intercept wouldn't be in 2013.
POV-Ray provides another point of view on single threaded performance, here the situation looks far less dire than under Cinebench:
Unfortunately the curves remain fairly distinct.
Once again, when we increase thread count we see AMD pull ahead.
SYSMark is a particularly telling benchmark as it is lightly threaded and does a good job of simulating all types of workloads:
The result here is AMD closing in, albeit slowly, on Intel's performance advantage. I suspect this is quite possibly the best case scenario for AMD, it doesn't necessarily want to surpass Intel in performance but it wants to get close enough where pricing and other factors (e.g. GPU performance in its APU parts) can make a bigger difference.
Our Visual Studio 2012 test is a good combination of single threaded and multithreaded workloads in one:
With Vishera, AMD did a lot to close the gap betwen itself and Intel. Another increase like this and we won't see AMD surpass Intel, but the two should remain fairly close.
These last two tests show us the other side of the coin. If both AMD and Intel continue on their present tracks, what will happen in a test where AMD already does well today?
In areas where AMD holds a significant advantage, Haswell would need to deliver more than a 15% gain in performance at the same price point to catch up.
None of the results here are all that surprising. AMD remaining on its current course isn't enough to dramatically change its standings vs. Intel in another year. Vishera definitely cut into the performance delta, but the 2013 follow-up will have to do even more to really have an impact. Steamroller is far more focused on increasing IPC, however without a new process node it'll be difficult to demonstrate another gain in frequency like we see today with Vishera. I suspect the real chance for AMD to approach parity in many of these workloads will be with its 20nm architecture, perhaps based on Excavator in 2014.
Facebook
Twitter
RSS
250 Comments
View All Comments
CeriseCogburn - Tuesday, October 30, 2012 - link
Yes, also partially perception, since SB was such a massively unbelievable overclocker, so expectations were sky high.I'd like amd fan boy to give his quick amd win comparison on overlcocking Ivy vs whatever amd crap he wants opposite.
I more than suspect Ivy Brdige will whip the ever loving crap out the amd in the OC war there, so amd fanboy can go suck some eggs, right.
I mean that's the other thing these jerks do - they overlook an immense Intel success and claim 2500K oc power is not needed, then babble about " the next Intel disappointment " immediately afterward, not compared to their amd fanboy chip, but to Intel's greatest ever !
I mean the sick and twisted and cheating nature of the amd fanboy comments are literally unbelievable.
zilexa - Tuesday, October 23, 2012 - link
He writes "Brazos had a mild update" uhmm noo sorryy. Brazos 2 hasn't been released to market.. so we havent seen any update here. Same for Trinity although you can buy some Trinity laptops. Not much. But nothing on desktop.Spacecomber - Tuesday, October 23, 2012 - link
Since AMD seems to be pushing multi-threaded performance at this time, it would have been interesting to see how this is born out in a more multi-threaded game, such as the Battlefield series titles. I know that in benchmarking my six core 960T (with unlocked cores), I could see some performance advantage between running this CPU with 6 cores versus the default 4 cores playing BFBC2. I'm not saying that this is where AMD will outshine Intel by any means, but it would have been an interesting test case for comparison's sake.(I actually suspect that by the time you go from 6 cores to 8 cores, you will have run out of any significant advantage to being able to handle the extra threads.)
BellFamily7 - Tuesday, October 23, 2012 - link
I think in Final Words Anand should add: "And the 4 year old Intel i7-920 - what a chip that was/is!" It is startiling to see AMD barely keep-up with the venrable 920 a good four years on - and those first charts in this article are with the i7-920 at stock speed. The average enthusiast is running an i7-920 at 3.8 Ghz on air all day long and achieving performance on par - or better - than many of today's CPU's!Of course, there is one big downside - power. This is Intel's big story to me: the speed and power of an O/C'd i7-920 on one quarter (or less) the power. Cool!
Thanks for putting the old i7-920 in the mix - it shows just what a ground breaking design it was...and in many ways, still is.
Senti - Tuesday, October 23, 2012 - link
Indeed i7-920 is the most awesome CPU in those graphs considering its age and nice overclockability. If there was overclocked version of it graps would look pretty funny.I use i7-930 @ 4.1 for a long time now and just can't justify my itching urge for upgrade. More than that, it'll probably survive here for 2 more years until Haswell-EP as plain Haswell looks handicapped in terms of compute power in favor of iGPU and power draw. I do NOT need power-restricted desctop CPU – with power saving features it'll do fine on idle with any max TDP.
ClagMaster - Tuesday, October 23, 2012 - link
With comprehensive ECC at the same price this would make a good server or workstation chip.AMD needs to get a 22nm process going and start some serious architectural soul-searching.
bwcbwc - Tuesday, October 23, 2012 - link
I don't understand why you keep saying that the 6300 fails to beat Intel at it's price point for the multi-threaded tests. At $130-140, the 6300 is going up against the core i3's and the multi-threaded benchmarks show the 6300 beating the core i3. Seriously: what am I missing here?Rhezuss - Tuesday, October 23, 2012 - link
I'd have loved to see the PHenom II X4 980 BE or any X6s in the comparisons...nleksan - Tuesday, October 23, 2012 - link
I can't see this as anything but a "win" for AMD, although there are certainly some sad feelings lingering about as I read this article regarding the 15% employee layoff that recently occurred. The promise AMD made was 10-15% improvements in IPC, and we certainly got that; not only that, but at a lower price than the first generation, AND with some very promising overclocking potential based on the scaling shown in this article.However, I refuse to acknowledge these as "FX" chips. "FX" was the designation given to the very first CPU produced by AMD that outperformed Intel's best offerings by a significant margin, the Socket 940 FX-51 2.2ghz single-core CPU based on the Opteron version of the Athlon64 architecture. The reason for my petty "harrumph-ing" is that I own one of the very first FX-51 chips released (from the first batch of 1,000), purchased the day of release back in 2003 alongside an Asus SK8V motherboard with 2GB of Corsair XMS3200RE DDR-400 Reg/ECC Dual-Channel RAM, and which served admirably with its brother the X800XT-PE 256MB GDDR2, until its well-deserved retirement in 2009.
That chip was, and to this day is, my favorite CPU of all time. It was a quirky chip: a server-backbone (and consequent unusual socket choice), ahead-of-its-time 64bit architecture, record-setting bandwidth at a "low" 2.2Ghz while Intel was trying their darnedest to hit 4.0Ghz with the P4, no real options in terms of a future upgrade path, and its champagne-tastes that could only be satiated by incredibly expensive Registered memory. However, it was FAST as all get-out, ran nice and cool with a Thermaltake Silent Tower Plus, and had a good amount of overclocking headroom for the time (an extra ~200-280Mhz was common).
Oh, and it DEMOLISHED the Pentium IV Emergency Edition CPU's that came clocked 55% higher! Paired with the best video card the world had ever seen at the time, it was unstoppable, and I recall running 3dMark for the first time after the build was finished only to nearly poop myself, as this 2.2Ghz chip STOCK just out-performed every single Intel CPU on the charts outside of those OC'd with the help of LN2.
I am working right now to rebuild the rig, as I feel it is time for it to come out of retirement and have some fun again, and I want to see just what it really is capable of with some better cooling (extreme-air or decent water).
[SPOILER]I have already lapped the CPU and the block (I am amazed at how poorly the two mated before; the chip AND cooler were noticeably convex), and based on the flatness of each it will certainly be good for a few degrees; add in the magic of today's best TIM's (PK-1) compared to that of 2003, the wonders of modern computer fans via 2x 92mm 3800rpm 72cfm/6.5mmH2O fans doing push-pull, and an extra 3 intake fans feeding it fresh air.... It will be a fun way to bring a memory back to life :D
Plus, the X800XT-PE has been thoroughly prepped for overclocking, with a 6-heatpipe heatsink and 92x20mm (61.3cfm/3.9mmH2O) swapped for the stock unit and mated with PK-1, EnzoTech Pure-Copper VGA RAM-Sinks attached to all of the card's modules with PK-1 and less than a needle-tip's worth of superglue at two of the four corners of each, and the same for the MOSFET/VRMs on the card. Combined with a pair of 120mm 69cfm fans blowing air across it (mounted on the inner side of the HDD cage opposite an intake fan), an 80x15mm 28.3cfm fan mounted to blow air directly on the back of the card, a PCI slot blower fan pulling hot air from the card and exhausting it out the back, as well as an 80x25mm 48cfm fan mounted where the lower PCI brackets used to be exhausting air... I think it'll do just fine ;)
[/SPOILER]
However, I am not taking any sides in this "CPU WAR". The minute one company starts to seriously pull ahead, the competition is lost, and we ALL lose. Innovation will become scarce, people will become excited about 5% IPC improvements from generation to generation and fork out the money for the next "great thing" in the CPU world, not to mention the cost for the constantly-changing socket interface.
AMD has been in a bad way for some time now, pretty much since the Core processors from Intel began to overrun their Phenom lineup. Sure, they had some really amazing processors for the money, such as the Phenom II X4 965BE/980BE/960T and X6 1055/1090/1100T, but Intel was still the performance leader with their E8600, Q6600, and the many QX9xxxx processors that transitioned into the still-strong X58/1136 platform (with the 920/930/975X/990X standing out), and they have only gained traction since.
I am no fanboy, and I hate to get onto any enthusiast site and scroll through comments sections where pimply-faced, Cheetoh-encrusted, greasy-haired know-it-all loser's frantically type away in a "Heated Battle of 'Nuh-Uh's' and 'Yuh-HUH!'".
(that is called hyperbole)
Fortunately, at least for the most part, I don't see that here.
Perhaps we should all go out and buy one of these new chips, maybe for a build for a friend or family member, or a home-theater PC or whatever, but regardless of whether you "Bleed Red" or "Bleed Blue", both "sides" will win if AMD gets the money to truly devote enough resources to one-upping Intel, or more likely, coming close enough to scare them. When the competition is closest, only THEN do we see truly innovative and ground-breaking product launches; and at the current rate, we may be telling our grandchildren about how "once, a long time ago, there was a company.... a company named AMD".
For the record, I AM NOT in any way a Fanboy; I buy whatever gives me the best bang-for-my-buck. Fortunately, at my job I am the only "tech-y" person there so whenever there is an upgrade in someone's equipment, or even servers, I get the "old" stuff :D I have sold literally hundreds of CPU's off that I had no use for, but I kept the favorites or the highest-end in each category that I was able to get. However, many of them I purchased myself (Opteron/Xeon from work, the rest I bought 90% of).
Here's a list of processors I currently have in possession, in my house, in the best reverse-chronological order I can remember:
i7-3930K (24/7 4.6Ghz - Max 5.2Ghz), i7-3820QM, i7-2600K, 4x Xeon E7-8870 (got 8 for $2k from work, sold 4 @ $2k/ea and built a Bitcoin Miner that earned me ~6,700Mh/sec with 4x 5970's in CF-X; earned over 1200BC and cashed out when they peaked at ~$17/ea for a massive profit and eventually stopped mining), i5-2400, Xeon X5690, i5-2430M, Opteron 6180SE 12-core, Xeon W3690, Phenom II X6 1100T-BE, Phenom II X4 980-BE, Phenom II X4 960T-BE (built girlfriend a rig: best CPU for $$; unlocked to 6-core; hits 4.125Ghz 6C / 4.425Ghz 4C), Xeon X7460, Core2Quad Q9650, 4x Xeon X3380's, Opteron 8439SE, Xeon X5492, Core2Duo E8600 (from Optiplex 960, hits 4.5Ghz on air), Core2Duo T7400, Athlon II X4 640 (E0), Athlon II X4 650, Pentium Dual-Core T4400, Turion II N550 Athlon X2 7750BE, AMD FX-62 (3.25Ghz easy), Xeon X3230, Athlon X2 5200+, Opteron 890, Turion II Ultra M660, Athlon64 X2 6400+ BE, Opteron 185, Athlon64 X2 4800+, Opteron 856, Opteron 156, AMD FX-51 (24/7 2.45Ghz stock voltage), Opteron 144 (OC'd to ~2.6Ghz), Turion ML-44, Pentium 4-EE 3.46Ghz (could barely hit 3.5Ghz...junk), Pentium 4 3.2Ghz, Pentium 4 2.8Ghz (easily ran at 3.6Ghz on air 24/7 with +0.015V, awesome CPU!), Celeron Mobile 1.6Ghz, Pentium 4 2.4Ghz, Celeron 1.8Ghz, and plenty more....
***Have a set of 8x Xeon-EP E5-4650 8-core's coming when we upgrade again in January; they are upgrading the whole rack so I am getting, along with the chips: 3 total 4-CPU boards, 384GB of DDR3-1333 Reg/ECC, the entire cooling system, 16x LSI/Adaptec RAID Controller Cards (all PCI-e x8, support minimum 24x SAS 6Gbs drives, have between 1 and 4GB of Cache, and all have BBU's), 96x 150GB 15Krpm SAS6 + 48x 600GB 15Krpm SAS6 enterprise drives, and about two-dozen Nvidia "professional" cards (12x Tesla M2090's, 4x Tesla K10's that were used to evaluate platform, and 8 Quadro 6000's) all for $1900!!!!!!!! The supplier offered $2150 for "Trade-Up" but I am really good friends with the entire IT department (all 6 of them) and they offered them to me instead! FOLDING@HOME WILL BE SHOWN NO MERCY!
CeriseCogburn - Tuesday, October 30, 2012 - link
Ok idiot, I've had enough already.First of all, nice amd fanboy story. Before you go insane claiming over and over again it is not true, I want to point out to you your own words....
1. Intel has been dominating since core 2
2. Without amd competing there will be no innovation and tiny 5% will cost and arm and a leg and the idiots will be spending all their money buying it
Okay, let's take those 2 statements, and add in SANDY BRIDGE and it's amazing architectural jump.
Whoops !
There goes your sick as heck and fan boy theory.
Furthermore with your obviously supremely flawed BS above, you did your little amd fanboy promotion saying we should all go out and buy one of these amd chips for a family member or some upgrade - BLAH BLAH BLAH BLAH BLAH.
Let's add in your eccentric Fx chip story, your declaration it's your favorite cpu of all time, your bashing of Intel claiming only Ln2 could bring Intel close, and then your holy of holies the resurrection build...
OK ? Forgive any and all of us who don't buy your I am not an insane amd fanboy lines.
Look in the mirror, and face the dark side, let it flow through you Luke, you are and amd fanboy, and Intel will innovate and make absolutely amazing cpu's like the SB even when amd is slapping itself in the face and choking and dying ... feel the anger amd fanboy - amd is NOT NEEDED... let it flow through you amd fanboy, your journey to dark side is nearly complete...
When you kill your greatest FX cpu rebuild, you will have crossed over to the darkside !