Subject: Processors, Shows and Expos | September 10, 2013 - 02:31 PM | Ryan Shrout
Tagged: quark, Intel, idf 2013, idf
In a very interesting and surprising announcement at the first Intel Developer Forum keynote this morning, Intel's new CEO Brian Krzanich showed the first samples of Quark, a new SoC design that will enter into smaller devices that even Atom can reach.
Quark is the family name for the new line of SoCs that are open, synthesizable and support with industry standard software. An open SoC is simply one that will allow third-party IP integration with the processor cores while a synthesizable one can be moved and migrated to other production facilities as well. This opens up Intel to take Quark outside of its own fabrication facilities (though Krzanich said they would prefer not during Q&A) and allow partners to more easily integrate their own silicon with the Quark DNA. Intel had previously announced that Atom would be able to integrate with third-party IP but that seems to have been put on the back burner in favor of this.
Quark will not be an open core design in the same way that ARM's core can be, but instead Intel is opening up the interface fabric for direct connection to computing resources.
The Quark SoC is square in the middle
Krzanich showed off the chip on stage that is 1/5 the size of Atom and 1/10 the power levels of Atom (though I am not sure if we are referring to Clover Trail or Bay Trail for the comparison). That puts it in a class of products that only ARM-based designs have been able to reach until now and Intel displayed both reference systems and wearable designs.
UPDATE: Intel later clarified with me that the "1/5 size, 1/10 power" is for a Quark core against an Atom core at 22nm. It doesn't refer to the entire SoC package.
Intel hasn't yet told us what microarchitecture Quark is based on but if I were a betting man I would posit that it is related to the Silvermont designs we are looking at on Bay Trail but with a cut down feature set. Using any other existing design from Intel would result in higher than desired power consumption and die size levels but it could also be another ground up architecture as well.
I'll be poking around IDF for more information on Quark going forward but for now, it appears that Intel is firmly planting itself on a collision course with ARM and Qualcomm.
UPDATE 1: I did get some more information from Intel on the Quark SoC. It will be the first product based on the 14nm manufacturing process and is a 32-bit, single core, single thread chip based on a "Pentium ISA compatible CPU core." This confirms that it is an x86 processor though not exactly what CPU core it is based on. More soon!
Subject: Processors, Shows and Expos | September 10, 2013 - 11:02 AM | Ryan Shrout
Tagged: live blog, keynote, Intel, idf 2013, idf
UPDATE: You can see the replay of our live blog below from Day 1 of IDF but be sure you head over to the Day 2 Live Blog page to set a reminder! Join us on Wednesday at 9am PT / 12pm ET!!
Today is the beginning of the 2013 Intel Developer Forum in San Francisco! Join me at 9am PT for the first of three live blogs of the main Intel keynotes where we will learn what direction Intel is taking on many fronts!
Subject: Processors, Mobile | September 4, 2013 - 11:32 AM | Ryan Shrout
Tagged: Temash, ifa 2013, asus, APU, amd, a4-1200
The hits just keep coming from ASUS this morning with the announcement of a new ultraportable notebook with the ambiguous name of X102BA. Though the name might not be catchy the device itself is more interesting because of the hardware that is powering it, specifically an AMD Temash A4-1200 APU.
This marks one of the few highly visible systems being powered by the AMD Temash architecture and I will be very curious to its reception. The APU itself is a dual-core part that runs at 1.0 GHz with integrated Radeon HD 8180 graphics that is more than enough for a modest Windows 8 working environment. There is a quad-core variant of Temash available but ASUS decided to go with the dual-core option. If you need more information on the new architecture that AMD created for Kabini and Temash (based on Jaguar CPU cores and GCN GPU cores) then you should see our coverage from their announcement back in May.
The rest of the specifications are a bit more tame, including a 1366x768 10.1-in 10-point multi-touch screen, USB 3.0, 802.11n WIFI, bundled Microsoft Office Home and Student 2013 and a touted 2-second resume time.
Even though the battery life is only listed at 5 hours, the 2.4 pound weight makes the X102BA a very portable machine. Plus you can get it in Hot Pink!
Subject: Motherboards, Processors | September 3, 2013 - 06:19 PM | Ryan Shrout
Tagged: x79, P9X79 PRO, Ivy Bridge-E, Intel, i7-4960X, asus
If you read our Intel Core i7-4960X Ivy Bridge-E review posted earlier today, you likely saw our overclocking results. After publication I got contacted by ASUS asking why we didn't attempt to overclock our CPU sample with one of their updated motherboards. In truth we were unable to get any of the pre-release UEFI firmware updates to apply to our P9X79 Pro or Rampage IV Extreme motherboards.
Move on to this afternoon and we were finally able to patch up the v1.02 of the P9X79 Pro and tossed in the same Core i7-4960X sample we used in our initial story. What were the results?
Click to Enlarge
As you can see above we were able to overclock the processor to 4.413 GHz at UEFI set voltage of only 1.40v. Previously we were only reaching a 4.3 GHz overclock and even had to up the voltage a bit higher.
I was hoping that I would be able to reach the 45x multiplier but alas it wasn't meant to be. I will keep messing with our 4960X to see how much further can push it.
Subject: Processors | September 3, 2013 - 05:43 PM | Jeremy Hellstrom
Tagged: 4960x, core i7-4960x, i7-4960X, Intel, Ivy Bridge-E, lga 2011, x79
You won't see the release of Intel's new processor as being described as "fascinating as whatever was happening with that rancher dude in Wyoming with the chickens and the laser pointer", you will have to head to The Tech Report to enjoy that type of comment. Nor will you finally learn that 5% of people who buy this chip "Need more knobs for extreme overclocking."; unfortunately he is probably right on the money as there are very few reasons to upgrade from Sandy Bridge-E to IVB-E. Stick your tongue in your cheek and read the usual benchmarks delivered a few percentage points faster than the last generation.
The truly masochistic can immediately follow that up with Ryan's review here.
"The NSA intercepted our review of the Core i7-4960X before we even had it completed. Let's listen in and see what they made of it."
Here are some more Processor articles from around the web:
- Intel Core i7 4960X Review @ OCC
- Intel Core i7 4960X @ AnandTech
- Intel Core i7 4960X Ivy Bridge Extreme Processor Review @ Hi Tech Legion
- Intel i7-4960X Ivy Bridge-E @ LanOC Reviews
- Intel Core i7-4960X Extreme Edition @ Bjorn3D
- Intel i7-4960X Ivy Bridge-E Review @ Hardware Canucks
- Intel Core i7 4960X, 4930K and 4820K tested @ Hardware.Info
- Intel Core i7 4960X EE CPU / Asus X79-Deluxe Motherboard @ Kitguru
- Intel Core i7-4960X Extreme Edition @ Legion Hardware
- Intel Core i7-4960X Ivy Bridge-E Review: New flagship, old flagship @ Techspot
Very Minor Changes
November 14th, 2011 - that is the date that Intel introduced the LGA 2011 socket and the Sandy Bridge-E processor. Intel continued their pattern of modifying their mainstream architecture, Sandy Bridge at the time, into a higher performance (and higher priced) enthusiast class. The new socket differentiated these components into their own category for workstation users and others who demand top performance. Today Intel officially unveils the Ivy Bridge-E platform with essentially the same mindset.
The top end offering under the IVB-E name is the Core i7-4960X, a six-core, HyperThreaded processor with Turbo Boost technology and up to 15MB of L3 cache. Sound familiar? It should. There is really very little different about the new 4960X when compared to the Sandy Bridge-E Core i7-3960X released in 2011. In fact, the new processors use the exact same socket and will work on the same X79 motherboards already on the market. (Pending, of course, on whether your manufacturer has updated the UEFI/Firmware accordingly.)
The Ivy Bridge-E Platform
Even though the platform and features are nearly identical between Sandy Bridge-E and Ivy Bridge-E there are some readers that might need a refresher or maybe had never really investigated Socket 2011 products before today. I'll step through the major building blocks of the new Core i7-4960X just in case.
Subject: Processors | August 30, 2013 - 04:28 PM | Jeremy Hellstrom
Tagged: x79, lga 2011, Ivy Bridge-E, i7-4960X
The i7-4960X has arrived and the fact that it is compatible with current LGA2011 boards might be the biggest hurdle for the Intel sales team. [H]ard|OCP tested it on a brand new ASUS X79 Deluxe and while it did prove to be a bit faster than a 3930K, or for that matter a 4770K, as well as using a little less power at full load it just does not offer enough of a jump to make swapping your SB-E chip out. Idle power is impressively low and if you are on an older LGA 1366 board you will certainly notice a jump, so there will certainly be a market for this generation of Intel chip.
"We debut Intel's next $1000 Extreme Desktop processor, the Core i7-4960X, this time with Ivy Bridge architecture and a couple of extra cores thrown in for good measure. It is a beast of a CPU for those that can actually harness its power and bandwidth, but how much better is it than Sandy Bridge-E and Haswell at the same clocks?"
Here are some more Processor articles from around the web:
- Intel Haswell Linux Performance Remains Mixed Against Windows @ Phoronix
- Intel Core i5-4430 CPU LGA1150 Haswell @ Benchmark Reviews
- Intel Haswell i5-4670K vs. i7-4770K Comparison @ techPowerUp
- Desktop CPU Comparison Guide @ TechARP
- The Workstation & Server CPU Comparison @ TechARP
- Mobile CPU Comparison Guide @ TechARP
Subject: Processors | August 23, 2013 - 03:28 PM | Ryan Shrout
Tagged: piledriver, FX-9590, amd
Over the last couple of days we had heard rumors about a potential price drop on the 5.0 GHz (Turbo Speed) AMD FX-9590 processor that was released in June. As the week progressed, the likelihood of this being true skyrocketed as several online outlets are showing much lower than expected pricing on the 8-core 220 watt CPU.
Two different UK-based online and retail outlets were showing the FX-9590 for sale for as low as £279 or $434 USD. That is a big price drop from £699 rate ($1008 USD) and obviously is causing quite a stir in the community. This puts the latest entries in the world of AMD FX just above the other parts like the standard FX-8350 in terms of cost which was definitely NOT the case in June or July.
In the US, the FX-9590 is still selling at Newegg as an OEM part but the current price is stuck at $879!
So why all the fuss? AMD claims that these are NOT price drops at AMD's request and that instead are the result of "business to business" negotiations. The official statement from AMD is as follows:
AMD channel partners are able to deliver the AMD FX-9000-series processor, AMD’s fastest and most powerful desktop processor, in highly customized systems and solutions in a manner that provides AMD fans access to the technology. We are excited to see high levels of interest in our AMD FX 9000-series processors, and will continue to work with our valued channel partners to ensure our products are readily available to the enthusiast community.
If you're like me, you don't really take anything interesting away from this statement other than "no comment." So what is really happening?
First, according to AMD the FX-9590 was never intended to be sold as an OEM part and rather was supposed to ship only in pre-built systems from companies like iBuyPower or in bundles that include a motherboard and cooler along with the processor. If these bundles were slow sellers though it seems plausible that the retailers would find ways to expire the bundle program and "accidentally" start selling the processors alone. Based on photos from ReviewBros that appears to be the case.
Photo source: ReviewBros
In reality though, this is the pricing that we would have liked to see the FX-9590 ship at originally and the first sets of reviews (that we were not included on) might have been much more positive. At $430 USD the FX-9590 competes with the higher end Core i7 Haswell processors in terms of performance but obviously uses quite a bit more power to get there.
If you are interested in buying a bundle or a system with the FX-9590 I do expect there to be some updates to pricing from all of the same system builders that launched with the processor originally to reflect these "business to business" happenings. I have already expressed interest to AMD and a couple of boutique builders in reviewing a system with these pricing and placement adjustments.
As for the idea of a "price drop", things are just more complicated than that. AMD tells me that because it was never intended to sell as an OEM part any pricing changes are not a result of AMD's demands. Honestly I don't know why AMD is so opposed to just saying there has been a price drop other than the negative reaction of the initial launch buyers; but that is always the case in the enthusiast market.
Regardless of the verbiage, the fact is that you'll likely be able to find the AMD FX-9590 and its 5.0 GHz Turbo clock rate available at lower prices in systems and on store shelves (though without AMD's consent) for much closer to the actual performance/value they offer. I'll take it.
Subject: General Tech, Graphics Cards, Processors, Mobile | August 23, 2013 - 01:31 AM | Scott Michaud
Tagged: Futuremark, AnTuTu, benchmarking
VR-Zone tossed the bees nest in a paint shaker and received a fairly sedate outcome.
A little background information is required. AnTuTu, a mobile benchmark developed by AnTuTu Labs, has been accused of inaccurate scores and bias towards specific hardware. Leaked BayTrail-T benchmarks, surpassing our expectations of Intel's capabilities, were harshly refuted based on AnTuTu's credibility. More recently, certain Samsung GPUs have been allegedly recorded self-overclocking during that benchmark but not elsewhere.
Scene from Cloud Gate, latest Windows 3DMark.
Oliver Baltuch, president of Futuremark, accepted an interview with VR-Zone to discuss business and ethics in their marketplace. Futuremark is a direct competitor to AnTuTu and a household name in the benchmarking community. Being modest Fins, self-proclaimed, they did not wish to discuss whether AnTuTu was less honest than they are. Futuremark does disagree with AnTuTu's process, however, and has some suggestions for better results.
The design process for 3D Mark Android begun with 25 pages of specification proposal. Each vendor is given a chance to reply to that proposal and these responses are compared. Changes to the specification must be reviewed by a committee sitting between the financial department and the engineering department.
Baltuch made the point that all of their finances for the last five years, according to Finnish law, can be reviewed for about $7 USD. Despite being a private company, the law mandates no deals can be made in secret.
On the engineering side of things, drivers are approved only if they follow specific guidelines. Unapproved results will be removed from their website and leaderboards followed by a polite conversation with the manufacturer. Drivers are not allowed to identify their benchmarks intent on modifying settings due to that information.
Almost every benchmark they release gets negative responses from some upset vendor or vendors.
The relatively short interview is wrapped up with commentary on iOS benchmarks. Futuremark is nearing completion of their first benchmarking app. Apple disallows apps to exceed 60 frames per second, through vsync, which unnecessarily hinders benchmark scores. Working around this, Futuremark developed a method to render frames which are not displayed on screen to keep the processors from idling once at frame rate cap.
Ryan must love that idea...
This concept has, according to the interview, reached internal QA review and is expected to be released in a few weeks.
Futuremark develops benchmarks for x86 Windows, Windows RT, Android, and iOS. Scores are intended to scale linearly to their metrics and are designed to allow cross-platform performance comparisons.
Subject: General Tech, Graphics Cards, Processors | August 22, 2013 - 01:39 PM | Ryan Shrout
Tagged: sony, ps4, playstation 4, Kabini, hUMA, amd
UPDATE: I have added new info at the bottom of this post with more commentary from AMD (kind of).
You might have seen some reports in the last couple of days claiming that the upcoming Sony PlayStation 4 (PS4) will have a big advantage over the Xbox One thanks to its unique ability to support AMD's hUMA memory architecture. hUMA, heterogeneous unified memory architecture, is an exciting new memory technology that AMD has built into upcoming APUs.
Josh published a story on hUMA that sums it as so:
The idea behind hUMA is quite simple; the CPU and GPU share memory resources, they are able to use pointers to access data that has been processed by either one or the other, and the GPU can take page faults and not rely only on page locked memory. Memory in this case is bi-directionally coherent, so coherency issues with data in caches which are later written to main memory will not cause excessive waits for either the CPU or GPU to utilize data that has been changed in cache, but not yet written to main memory.
There's just one problem with these various reports (VR-Zone, ExtremeTech): they're incorrect. After sending some emails to our representatives at AMD I was told that "Kabini doesn't support hUMA" which is the APU that both the PS4 and Xbox One processors are based on. AMD further clarified with us:
Our spokesperson made inaccurate statements about our semi-custom APU architectures and does not speak for Microsoft, Sony or the AMD semi-custom business unit responsible for co-developing the next generation console APUs.
So while the PS4 will still be a faster system thanks to its higher SIMD processor (GPU core) count, there is no support for a true heterogeneous unified memory architecture in either upcoming console platform.
NOTE: I have had several people point out that it's possible Sony and Microsoft worked on their own custom memory architectures that will perform similar functionally to hUMA. That is entirely possible but means that official hUMA support isn't on the SoCs.
UPDATE: AMD contacted me again to make another comment. Essentially, they said that the correction statement to the original statement claiming hUMA was part PS4 was "inaccurrate" but that this correction does NOT mean the opposite claim is true. Even when pressed for a more specific and debate-ending comment, AMD wouldn't give us any more information.
So does the PS4 have support for some type of heterogeneous unified memory? Maybe. And the Xbox One? Maybe. At this point, I'd stop listening to anything AMD has to say on the subject as they are likely to recant it shortly thereafter. Many readers have emailed me with their thoughts and I personally feel that its more likely the original statement from AMD (that the PS4 will have the edge with a hUMA design) will turn out to be the truth in the long run...