Ah, the end of August. School is about to start. American college football is about to get underway. Hot Chips is now in full swing. I guess the end of August caters to all sorts of people. For the people who are most interested in Hot Chips, the amount of information on next generation CPU architectures is something to really look forward to. AMD is taking this opportunity to give us a few tantalizing bits of information about their next generation Steamroller core which will be introduced with the codenamed “Kaveri” APU due out in 2013.
AMD is seemingly on the brink of releasing the latest architectural update with Vishera. This is a Piledriver+ based CPU that will find its way into AM3+ sockets. On the server side it is expected that the Abu Dhabi processors will also be released in a late September timeframe. Trinity was the first example of a Piledriver based product, and it showed markedly improved thermals as compared to previous Bulldozer based products, and featured a nice little bump in IPC in both single and multi-threaded applications. Vishera and Abu Dhabi look to be Piledriver+, which essentially means that there are a few more tweaks in the design that *should* allow it to go faster per clock than Trinity. There have been a few performance leaks so far, but nothing that has been concrete (or has shown final production-ready silicon).
Until that time when Vishera and its ilk are released, AMD is teasing us with some Steamroller information. This presentation is featured at Hotchips today (August 28). It is a very general overview of improvements, but very few details about how AMD is achieving increased performance with this next gen architecture are given. So with that, I will dive into what information we have.
Less Risk, Faster Product Development and Introduction
There have been quite a few articles lately about the upcoming Bulldozer refresh from AMD, but a lot of the information that they have posted is not new. I have put together a few things that seem to have escaped a lot of these articles, and shine a light on what I consider the most important aspects of these upcoming releases. The positive thing that most of these articles have achieved is increasing interest in AMD’s upcoming products, and what they might do for that company and the industry in general.
The original FX-8150 hopefully will only be a slightly embarrasing memory for AMD come Q3/Q4 of this year.
The current Bulldozer architecture that powers the AMD FX series of processors is not exactly an optimal solution. It works, and seems to do fine, but it does not surpass the performance of the previous generation Phenom II X6 series of chips in any meaningful way. Let us not mention how it compares to Intel’s Sandy Bridge and Ivy Bridge products. It is not that the design is inherently flawed or bad, but rather that it was a unique avenue of thought that was not completely optimized. The train of thought is that AMD seems to have given up on the high single threaded performance that Intel has excelled at for some time. Instead they are going for good single threaded performance, and outstanding multi-threaded performance. To achieve this they had to rethink how to essentially make the processor as wide as possible, keep the die size and TDP down to reasonable sizes, and still achieve a decent amount of performance in single threaded applications.
Bulldozer was meant to address this idea, and its success is debatable. The processor works, it shows up as an eight logical core processor, and it seems to scale well with multi-threading. The problem, as stated before, is that it does not perform like a next generation part. In fact, it is often compared to Intel’s Prescott, which was a larger chip on a smaller process than the previous Northwood processor, but did not outperform the earlier part in any meaningful way (except in heat production). The difference between Intel and AMD in this aspect is that as compared to Prescott, Bulldozer as an entirely new architecture as compared to the Prescott/Northwood lineage. AMD has radically changed the way it designs processors. Taking some lessons from the graphics arm of the company and their successful Radeon brand, AMD is applying that train of thought to processors.
Get Out the Microscope
AMD announced their Q1 2012 earnings last week, which turned out better than the previous numbers suggested. The bad news is that they posted a net loss of $590 million. That does sound pretty bad considering that their gross revenue was $1.59 billion, but there is more to the story than meets the eye. Of course, there are thoughts of “those spendthrift executives are burying AMD again”, but this is not the case. The loss lays squarely on the GLOBALFOUNDRIES equity and wafer agreements that have totally been retooled.
To get a good idea of where AMD stands in Q1, and for the rest of this year, we need to see how all these numbers actually get sorted out. Gross revenue is down 6% from the quarter before, which is expected due to seasonal pressures. This is right in line with Intel’s seasonal downturn, and in ways AMD was affected slightly less than their larger competitor. They are down around 2% from last year’s quarter, and part of that can be attributed to the continuing hard drive shortage that continued to affect the previous quarter.
More MHz for the Masses
AMD has had a rough time of it lately when it comes to CPUs. Early last year when we saw the performance of the low power Bobcat architecture, we thought 2011 would be a breakout year for AMD. Bulldozer was on the horizon and it promised performance a step above what Intel could offer. This harkened back to the heady days of the original Athlon and Athlon 64 where AMD held a performance advantage over all of Intel’s parts. On the graphics side AMD had just released the 6000 series of chips, all of which came close in performance to NVIDIA’s Fermi architecture, but had a decided advantage in terms of die size and power consumption. Then the doubts started to roll in around the April timeframe. Whispers hinted that Bulldozer was delayed, and not only was it delayed it was not meeting performance expectations.
The introduction of the first Llano products did not help things. The “improved” CPU performance was less than expected, even though the GPU portion was class leading. The manufacturing issues we saw with Llano did not bode well for AMD or the upcoming Bulldozer products. GLOBALFOUNDRIES was simply not able to achieve good yields on these new 32 nm products. Then of course the hammer struck. Bulldozer was released, well behind schedule, and with performance that barely rose above that of the previous Phenom II series of chips. The top end FX-8150 was competitive with the previous Phenom II X6 1100T, but it paled in comparison to the Intel i7 2600 which was right around the same price range.
Subject: Processors | April 3, 2012 - 12:43 PM | Jeremy Hellstrom
Tagged: virtualization, ubuntu 12.04, Sandy Bridge E, Intel, FX 8150, Core i7 3960X, bulldozer, amd
Phoronix is taking the latest Ubuntu release and testing the performance on AMD's FX 8150 against Intel's Core-i7 3960X to see their relative performance in a virtual environment. Both machines had issues, Xen had critical issues which prevented it from running on the Bulldozer and ASUS motherboard system, while the Sandy Bridge chip had issues with Virtualbox. The testing was not so much a comparison of the performance difference between the two chips as it is a test of efficiency of these processors running tasks when virtualized. As both chips averaged 90%+ of base performance when virtualized you can see that both architectures have come a long way in this particular usage.
Also, keep your eyes out for a CPU review from Josh which should be arriving soon.
"With the upcoming availability of Ubuntu 12.04 "Precise Pangolin" being a Long-Term Support (LTS) release that will be quickly making its way into many enterprise environments, here's a look at the virtualization performance of this popular Linux distribution. In particular, being looked at is the Linux virtualization performance of KVM, Xen, and Oracle VirtualBox compared to bare metal when using Intel Sandy Bridge Extreme and AMD Bulldozer hardware."
Here are some more Processor articles from around the web:
- Xeon E5 2600: Interview with Intel IT's Ajay Chandramouly @ TechSpot
- Intel To Launch Ivy Bridge Desktop Processors This Week @ TechARP
- Mobile CPU Comparison Guide @ TechARP
- Intel Ivy Bridge Overclocking with the Core i7 3770K and Core i5 3570K CPUs @ Tweaktown
- AMD A8-3870 FM1 CPU @ Rbmods
Subject: Processors | February 28, 2012 - 12:51 PM | Josh Walrath
Tagged: trinity, FX-8120, FX-6200, FX-4170, FX, FM3, bulldozer, amd, am3+
Since AMD held their Analysts’ Day, we have not heard a whole bunch from their CPU division. The graphics side has been in full gear launching the HD 7000 series of products, and soon we will see the final pieces of that particular puzzle fall into place. What about the CPU group? We have heard about Trinity for ages now, but that particular launch is still months away. The last CPU update detailed the “K” series of unlocked Llano chips. What about Bulldozer? Is there a new stepping? How is GLOBALFOUNDRIES’ 32 nm SOI/HKMG progressing?
I don’t have all those answers, unfortunately. Since AMD proceeded to sack most of the PR team, our contacts have all but disappeared. Questions emailed to AMD are often not returned. Requests for CPU information (or samples) are ignored. Are these people just simply overworked, or is AMD clamping down on information? Hard to say. My guess here is that they are taking the philosophy of, “No news is good news.” If a company does not send out review samples, they do not have to deal with products receiving bad reviews. I am not saying that the FX processors are necessarily bad, but they do not match up well against Intel’s latest Sandy Bridge parts. At least AMD parts are priced appropriately overall for their level of performance. If we look at overall results, the FX-8150 does match up fairly well with the i5-2500K, and they both exist very close to each other in price points.
What we do know is that AMD has released two new processors into the market with the FX-4170 and the FX-6200. The FX-4170 is a new dual module (four core) 125 watt TDP part that is clocked at an amazing 4.2 GHz stock speed, and a turbo that goes to 4.3 GHz. This is the fastest consumer grade processor in terms of clockspeed, but it obviously is not the fastest processor on Earth. The original FX-4100 is a 95 watt TDP part at 3.6 GHz stock/3.8 GHz turbo, 4 MB L2 cache, and 8 MB of L3. The FX-6200 is perhaps the more interesting of the two. It has a base clock of 3.8 GHz and a max turbo speed of 4.1 GHz. This is a pretty hefty increase from the FX-6100 with its base 3.3 GHz and 3.9 GHz turbo. The 6100 is a 95 watt TDP part while the new 6200 is 125 watt TDP. The 6200 is a three module (six core) part with 6 MB of L2 cache and 8 MB of L3.
The last bit of news is that the FX-8120 is getting a price cut to put it more in line against the competition. The email that we received about this and the previous announcements was amazingly generic and fairly uninformative. We do not know the prices, we do not know the rollout schedule, and we have no idea how much the FX-8120 is going to be chopped. We have seen the retail market already cut the prices down on the FX-8xxx series. The high end FX-8150 was introduced around $289 but now it can be readily available for $259. Now that demand has dropped in the PC sector and AMD’s supply has caught up, it is no wonder we are seeing new SKUs and the lowering of prices.
My goal is to try to get a hold of some of these parts, as they do look interesting from a value standpoint. The FX-6200 is of great interest for many users due to the nice provisioning of cores, L3 caches, and speeds. Throw in a decent price for this particular product, and it could be a favorite for budget enthusiasts who want to stick with AMD products. The area where it does fall down is that of TDP when compared to Intel’s Sandy Bridge parts at that price point. The jump to 3.8 GHz base speed and 4.1 GHz turbo should make it very comparable in stock clocked performance to anything Intel has in that price range.
Overclocking could be interesting here, but since it is already a 125 watt TDP part I do not know how much headroom these products have. 4.8 GHz is very likely, but on air cooling I would not expect overclocked speeds to reach much more above that. Still, these are interesting parts and give plenty of bang for their price. Add in pretty mature support for AM3+ motherboards, and AMD still has a chance with enthusiasts. The only real issue that is looming is PCI-E 3.0 support for the AM3+ ecosystem. We have not heard anything about the upcoming (or is it cancelled?) 1090FX chipset, other than it is based on 890FX/990FX and should not support PCI-E 3.0. With AMD’s push for APUs, I would expect the upcoming Trinity parts to introduce PCI-E 3.0. AMD also looks like they will start funneling the enthusiasts towards FM2 platforms and Trinity based parts. While AMD looks to support AM3+ with Piledriver based cores, my best guess here is that AM3+ will be phased out sooner rather than later.
The next 6 months will be critical for AMD and their path moving forwards. At the very least we will have a better idea of where the company is going under the new management. I am still expecting some big changes from AMD, and if Trinity can give Intel a run for its money in terms of per clock CPU performance, then they could have a winner on their hands and adjust their roadmap to further exploit that particular product release.
Subject: Processors | February 17, 2012 - 01:52 PM | Jeremy Hellstrom
Tagged: fx-8150, FX, cpu, bulldozer, amd, 990fx
AMD's $270 flagship processor, the 3.6GHz FX-8150 had a mixed reception as the hype which lead up to the release built up our expectations to a point that the processor could not live up to. Part of the disappointment has been blamed on the Windows 7 thread scheduler, which AMD described as not being optimized for their architecture, which lead to the release of hotfix files KB2645594 and KB2646060. TechPowerUp revisited their benchmarks to see if these patches effectively increase the performance of multi-threaded tasks; single threaded tasks are dependant on processor speed so they should be unaffected by the patches.
"After settling on the market, with all the quirks and bugs supposedly fixed, all the hype and disappointment blown away, we put AMD's FX-8150 under the scope. Benchmarks are done with and without the Windows 7 hotfix and in depth overclocking should resolve any doubts you have about AMD's flagship processor."
Here are some more Processor articles from around the web:
- AMD A8-3870K and Sapphire HD6450 FleX @ Kitguru
- The Opteron 6276: a closer look @ AnandTech
- AMD A8-3870K Unlocked Llano APU Review @ Hardware Canucks
- The Workstation & Server CPU Comparison Guide @ TechARP
- Intel Core i7-3820 vs. Core i7-2700K and Core i7-3930K @ X-bit Labs
- Intel Core i7-3820 3.6GHz Processor Review @ Legit Reviews
- Intel Core i7-3820 @ Techspot
- Intel Core i7-3930K @ OC3D
Subject: Systems | February 1, 2012 - 04:10 PM | Jeremy Hellstrom
Tagged: amd, interlagos, bulldozer, supermicro, opteron 6200
Over at The Inquirer you can take a look at the performance of the Opteron 6274 as a server chip, as opposed to the desktop benchmarks that have made up the bulk of Bulldozer reviews on the web. SuperMicro has assembled a server containing a dual-socket Opteron 6274 for a total of 32 cores and 64GB of ECC DDR3-1333 RAM across eight channels running on 64-bit Windows Server 2008 R2 SP1. While the machine was quite powerful it still has difficulty keeping up with Intel, for instance its performance on CineBench was about the same as provided by an X5680 Westmere Xeon which is not Intel's best silicon. On the plus side, the scaling for multithreaded applications was quite good.
"IS BULLDOZER better in a workstation than a desktop? Let's see if it can be.
AMD's Bulldozer chip, while eagerly awaited, didn't really have a stellar debut as a mainstream - or high end, for that matter - desktop processor, still having a way to go before seriously challenging the incumbent Intel. The core, cache and memory performance all need a bit more power, although recent news such as Microsoft Windows' kernel patches do seem to help a little bit towards extracting more oomph from the unusual 'two integer units sharing one floating-point' approach."
Here are some more Systems articles from around the web:
- Alienware X51 Launch Details @ Hardware Heaven
- ZOTAC ZBOX AD04 Plus AMD Fusion Mini PC Review @ HardwareHeaven
- Sony Z Series @ Kitguru
- Ultimate SFF Lanparty Machine: Intel Xeon E5 (SNB-EP 8C/16T) on an ASUS Rampage IV GENE @ VR-Zone
Subject: Processors | January 30, 2012 - 04:51 PM | Jeremy Hellstrom
Tagged: Intel, amd, sandybridge, llano, bulldozer, Sandy Bridge E
If you are looking for a quick way to contrast the processors that were released this year then iXBT Labs has a review for you. They've added their CPU/APU reviews for the past year together and compiled some rather lengthy charts which reflect the comparative performance of a few older chips as well as the majority of chips released this year. Both Intel and AMD desktop and server chips are included, mobile users will need to look elsewhere to compare chips designed specifically for laptops. Their benchmarks range from 3D modelling to 3D gaming as well as compression, office suites and raster graphics processing so no matter what purpose you will be putting these chips to you should be able to get an idea what chips to be on the look out for.
Told you it was big, visit iXBT Labs if you want the readable version.
"The year 2011 has ended, so it's high time to sum up the results and see the general picture. If you're looking to upgrade, we hope this will make choosing a processor easier."
Here are some more Processor articles from around the web:
- The AMD FX (Bulldozer) Scheduling Hotfixes Tested @ AnandTech
- Intel Core i7-3930K vs Core i7-3820 vs FX-8150 vs 990X vs 2700K Review @ HardwareHeaven
- AMD Llano A8-3870K APU Review @ TechwareLabs
- AMD A8-3870K Unlocked Llano 3.0 GHz Quad Core APU Review @ Hi Tech Legion
- AMD A6 3500 triple core APU @ Guru of 3D
- AMD A8-3870K CPU Review @ Neoseeker
- CPU Performance Comparison Guide @ TechARP
- Intel Core i7-2600K 3.40GHz Processor w/Turbo Boost Technology Long Term Review @ ModSynergy
Q4-2012 In a Nutshell
Tis the reporting season. Yes, that time of year when some of the major players in the computing world get together and tell us all how well they did this past quarter. Ok, so they do not necessarily get together to announce results, but they sure time them that way. Today was AMD’s turn (and Apple’s), and the results were not nearly as positive as what Intel had to offer a few days ago.
Q4 2011 was flat in terms of revenue as compared to Q3. The company had gross revenue of $1.69 billion and had a net income loss of $177 million. That net income is not necessarily a bad result, but more on that later. Margins rose to 46%, which is still a far cry from Intel’s 65% for the past quarter. Gross revenue was up 2% from last year, which considering the marketplace and Intel’s dominance, is a solid win for AMD.
When we start talking about non-GAAP results, AMD had a net income of $138 million. The difference between those two numbers (a loss vs. a nice profit) is that the loss came from one time writeoffs. AMD has lowered its stake in GLOBALFOUNDRIES to 8.8%, and in so doing incurred a hefty charge. This is not so much money lost as it is lost value in the company.