In a final hurrah for AMD’s Bulldozer and its derivatives, Bristol Ridge APUs will launch later this year. Coming in just before Zen arrives in Q4, the update will bring Excavator to the desktop as well as introduce the new Socket AM4. Today, we’ve been treated to the Geekbench scores for the FX 9800P. Given the results, it looks like Excavator will be a nice IPC increase over the current Steamroller APUs.
Deviating from the rumoured 2.7Ghz base clock, the FX 9800P in the Lenovo 59AC sample runs at 1.85Ghz. Of course, this could be off given that Geekbench might not be properly reading the clock speed. The chip managed to score 2216 in the single-threaded tests and 5596 in the multi-threaded portion. This is pretty competitive compared to the FX 8800P especially given the clock speeds. Given that we don’t know the cTDP setting, we can’t draw too many conclusions.
The biggest change compared to Carrizo is the use of AM4, FP4 and DDR4. Bristol Ridge will showcase the motherboards and memory controller that Zen will be using and that will be the most interesting part about it. By finally bringing Carrizo to the desktop in numbers, AMD will have a new desktop architecture since 2015.
Just as the final versions of Oculus Rift and HTC Vive have been revealed and launched, another contender has jumped into the VR scene with their own solution. Unlike their competitors, though, Toronto-based Sulon has managed to provide a solution that won’t break the bank. Dubbed a rather plain Sulon Q, the new VR HMD requires no host system to use, relying completely on its own internal inertial tracking and AMD PC grade hardware, all strapped to your head.
Reading the spec sheet of the Sulon Q makes it seem like you’re strapping a Windows PC with a VR display rather than a VR display with an attached PC. The Sulon Q comes with a quad-core AMD FX-8800P mobile APU with 512 GCN shader cores, enough to give Intel’s mainstream mobile CPUs a run for their money. Combined with a 256GB SSD, 8GB of RAM, and a 2560×1440 OLED display with an 110° field of view, the entire system is strapped securely to the user’s face.
This Windows 10 PC might seem a bit underpowered compared to the hefty 290/390/970 GPU requires of the Rift/Vive but the console-like hardware standardization and DX12 should provide enough for a portable VR experience. Tracking is done by a Spatial Processing Unit with two front cameras which removes the need for external trackers and hardware. This means it will be easy to setup and use right away.
There is even gesture control available through the 2 front cameras as well, removing the need for a controller in some cases. This also allows for augmented reality, putting it into the same realm as Hololens. If might even be possible to move around freely with this display as the cameras can provide a feed of the surroundings, avoid awkward issues like bumping into people and things.
From the demo shown off at Capsaicin, it looks like rather than targeting gaming audiences which require lots of action/FPS and visual fidelity, the Sulon Q is meant to provide a first VR home experience with lower graphics quality and simpler gameplay. A console-like approach means a common hardware platform for developers to target and DX12 should help things along nicely. Looking at the demo, the graphics quality looked like something from the PS2/3 era which is passable when it comes to gaming.
In working with Sulon, this is another one of AMD’s prongs to push VR adoption to the masses. If the Sulon Q can later be used to display from a PC, it may well be the perfect entry level VR device. With a late spring launch (within 2 months), we can hopefully get some more information before it debuts for real.
Despite being stuck on the 28nm node for the past while, AMD has been working to bring more CPUs to consumers with better binning and performance as the process and technology matures. Case in point is the A10-7860K Godavari and Athlon X8 845 Carrizo chips launching today. Both chips are focused on increasing performance/watt either by using a new architecture design or simply better binning and process improvements.
First up is the 2M/4T Athlon X4 845, a 65W Carrizo part based on of the “Excavator” architecture. As expected of an Athlon part, it does not have an iGPU and uses the FM2+ platform with DDR3 and PCIe 3.0 x8. As the sole desktop Excavator part, the chip is a bit of a curiosity with only 2MB of L2 instead of the usual 4MB but still manages to have a higher IPC due to improved pre-fetch, large L1 cache and better branch prediction. The low power nature of Excavator also limits the clock speeds to 3.5-3.8Ghz.
Next we have the A10-7860K which is a Steamroller based 2M/4T chip with 512 GCN Stream Processors. With a 3.6-4.0Ghz clock speed, the chip places lower than the 7890K and 7870K but it comes in with a 30W lower TDP at 65W. Despite the lower TDP, the 7860K will get the 95W cooler which should make it a good choice for an HTPC/budget gamer. Overall these chips should tide AMD over till more Excavator and Zen based APUs hit later this year.
Earlier this year, AMD launched the A-series APUs under the “Carrizo” codename which strive for energy efficiency and lower wattage demands. In 2014, AMD outlined the 25×20 energy strategy to produce chips 25 times more efficient than current products by 2020. According to AMD’s research team, the extremely efficient Carrizo architecture has put the company on course to reach its 2020 target. More specifically, Carrizo chips alter the core voltage to gauge power demands and ensures the maximum frequency is only used when required.
In the enthusiast market, AMD has struggled to compete with Intel especially in single-threaded performance. However, this is a fairly niche sector and it’s sensible for AMD to work hard to manufacture low-cost, high-yield APUs which provide an excellent wattage to performance ratio. In the future, discrete graphics cards might become obsolete and replaced by APUs as computational demands are offset to a server. Whatever the case, AMD needs to make their products more energy efficient and that also applies to the Radeon brand. Thankfully, the Fiji architecture is a step in the right direction and illustrates AMD’s policy towards modern CPUs and GPUs.
Despite this, the majority of press coverage will surround AMD’s future high-end desktop CPUs and I hope they can produce something to shake up the market and make Intel feel less comfortable.
Thank you Venturebeat for providing us with this information.
Earlier this week, we were invited to participate in a conference call with AMD. In the last few weeks, AMD has enticed into strict deadline shrouded calls that promised to show off the new technologies ready for the upcoming Summer show season; namely Computex and E3. As we already know, AMD are hosting a dedicated event at E3 to potentially unveil the new Radeon 300 series graphics cards, but what about the upcoming Godavari/ Kaveri Refresh and Carrizo APU’s?
Last week we took a call regarding the Godavari/ Kaveri Refresh APU series, being headed by the flagship A10-7870k. On Monday, we took a call regarding Carrizo APU line-up. Godavari is aimed for the desktop enthusiast market, where Carrizo will be notebook orientated. The notebook sector has a huge potential consumer base of around 63 million, with around 2/5’s of those consumers generally aiming to spend between $400-$700. The 6th generation (Carrizo) will be aimed to tackle new trends set by our current media requirements. Consumers have the pre-conception that desktops are the central hub of the home and the media services, however, in a recent survey it was found that notebooks were the second-largest platform for media-streaming with TV’s being first and tablets being third.
The 6th generation APU’s offers what most consumers are looking to achieve with a notebook.
The Carrizo processors feature a completely redesigned architecture to optimise power efficiency and provide more performance. This means that during gaming in huge titles such as DOTA2, League of Legends, etc.. you can see an increase in performance compared to the 5th generation APU’s and also have longer playing time.
Some key features that really interested me were the HEVC/ H.265 hardware decode, HSA 1.0 and hand gestures. Even though hand gestures have been featured on notebooks before, never have they been implemented at this price point. HSA 1.0 helps with AMDs Looking Glass, this feature scans through all discoverable pictures and videos and uses facial recognition to sort them into categories.
AMD is busy releasing the OEM models and mobile solutions in order to get that part out of the way early as well as prevent anything from stealing news space from their highly anticipated next generation of CPUs, APUs and not to forget GPUs such as the R9 390x liquid cooled micro card.
This time around they’re ready with five new Carrizo mobile APUs that already have begun shipping and being available in Greater China. Worldwide availability will follow shortly as the chips make their way across the globe to manufacturers and resellers. The information is relative sparse at the moment, but we do know that none of them have a TDP higher than 25W and go all the way down to 2W, have two to four CPU cores and a clock speed up to 2.5GHz backed by up to 2MB cache. The only exception to the quad-core and 2MB cache scheme is the small AMD E1-7010 APU that only has half of both.
The Carrizo platform will be extended to the desktop market later and it will also enter the FX processor line-up. But until that time arrives, AMD cut the prices on the current Kaveri desktop APUs. Cheaper products is a thing that we always like to see and AMD cuts up to a third of their original pricing.
Thank you TomsHardware for providing us with this information
AMD’s planned 2016 next-gen APUs are rumoured to be 16-core processors, based on the Zen architecture, featuring a Greenland GPU, quad-channel DDR4 support, and HBM memory, according to TweakTown.
The forthcoming APU will replace the Godaveri platform, first introduced with the Carrizo APU. AMD is releasing a series of Carrizo notebooks later this year, likely to be launched at Computex 2015 in June.
The Greenland GPU is expected to be based around the Fiji architecture, which powers AMD’s flagship graphics card, the Radeon R9 390X. Though the 390X uses HBM1, AMD is expected to switch to HBM2 for Greenland next year.
AMD revealed at the prestigious International Solid State Circuits Conference (ISSCC) that its upcoming A-Series Accelerated Processing Unit (APU), codenamed “Carrizo”, will deliver new, advanced power management technologies while achieving substantial performance through new “Excavator” x86 CPU cores and a new generation of AMD Radeon GPU cores. Using a true System-on-Chip (SoC) design, AMD expects Carrizo to reduce the power consumed by the x86 cores alone by 40 percent, while also providing substantial gains in CPU, graphics, and multimedia performance over the prior generation APU; perfect for its intended audience of low power computers and notebooks.
“As a part of our continued focus on building great products, the advanced power and performance optimisations we have designed into our upcoming ‘Carrizo’ APU will deliver the largest generational performance-per-watt gain ever for a mainstream AMD APU,” said Sam Naffziger, AMD Corporate Fellow and co-author of the AMD presentation at ISSCC. “There have been remarkable advances in performance and energy efficiency in computing since the birth of the modern microprocessor. However, the energy-related benefits that flow from new manufacturing processes have slowed, ushering in an era when alternative ways to improve processor performance and efficiency are needed. AMD has been pursuing Heterogeneous System Architecture (HSA) and proprietary power management technologies to make continued gains. The upcoming ’Carrizo’ APU takes a big step toward the AMD 25×20 energy efficiency goal and incorporates a wealth of new features that will be adopted across our full product line going forward.”
New Carrizo Disclosures at ISSCC:
29% more transistors in nearly the same die size as its predecessor, “Kaveri”;
New “Excavator” x86 cores provide an uplift in instructions-per-clock at 40% less power;
New Radeon GPU cores with dedicated power supply;
Dedicated, on-chip H.265 video decode;
Double digit percentage increases in both performance and battery life;
Integrated Southbridge for the first time on an AMD high-performance APU.
By using new high-density libraries design, AMD managed to squeeze 29% more transistors on Carrizon, 3.1 billion, in roughly the same size chip as Kaveri This density increase has allowed a larger area for graphics, multimedia offload, and integration of the “Southbridge” system controller on a single-chip. The increased support for multimedia includes the new, high-performance H.265 video standard and double the video compression engines of its predecessor. The inclusion of H.265 in hardware will support true 4K resolutions, help extend battery life, and reduce bandwidth requirements when viewing compatible video streams.
The additional transistor budget also allows Carrizo to become the first processor in the industry designed to be compliant with the HSA 1.0 specification. HSA makes programming accelerators such as the GPU far simpler, ideally leading to greater application performance at low power consumption. Chief among the design advantages for HSA is the heterogeneous Unified Memory Access (hUMA) within Carrizo. With hUMA, the CPU and GPU share the same memory address space. Both can access all the platform’s memory and allocate data to any location in the system’s memory space. This coherent-memory architecture greatly reduces the number of instructions required to complete many tasks, thus helping improve both performance and energy efficiency.
New Energy Efficient Features
Several new power efficient technologies make their debut on the Carrizo APU. To deal with transient drops in voltage, which is known as droop, traditional microprocessor designs supply excess voltage on the order of ten to fifteen percent to ensure the processor always has appropriate voltage. But over-voltage is costly in terms of energy because it wastes power at a rate that is proportional to the square of the voltage increase; i.e. 10% over-voltage means about 20% wasted power.
AMD has developed a number of technologies to optimise voltage. Its latest processors compare the average voltage to droops on the order of nanoseconds, or billionths of a second. Starting with the Carrizo APU, this voltage adaptive operation functions in both the CPU and the GPU. Since the frequency adjustments are done at the nanosecond level, there’s almost no compromise in computing performance, while power is cut by up to 10 percent on the GPU and up to 19% on the CPU. Another power technology that debuts in Carrizo is called adaptive voltage and frequency scaling (AVFS). This technology involves the implementation of unique, patented silicon speed capability sensors, and voltage sensors in addition to traditional temperature and power sensors. The speed and voltage sensors enable each individual APU to adapt to its particular silicon characteristics, platform behaviour, and operating environment. By adapting in real-time to these parameters, AVFS can lead to up to 30 percent power savings.
In addition to helping reduce power use on the CPU by shrinking the core area, AMD worked to optimise the 28nm technology for power efficiency, and tuned the GPU implementation for optimal operation in power-limited scenarios. This enables up to a 20% power reduction over the Kaveri graphics at the same frequency. Combined, AMD’s energy efficiency innovations aim to deliver power savings on the order of a manufacturing technology shrink while staying in a well-characterised, cost-optimised 28nm process.
According to the latest leak coming from notebook manufacturers, the new AMD Carrizo APU will make its début as early as December 2014. This however wont be the desktop version but rather the Carrizo-L APU. It is designed for the entry-level notebook segment and will replace the current Beema and Mullins APUs. The full version of Carrizo is said to be scheduled for March 2015.
The Carrizo-L will be set to compete against Intel’s entry-level Pentium and Celeron processors. The new APU is still built upon the 28nm process and has support for dual-channel DDR3-2133 memory. It isn’t really a surprise that AMD didn’t go for DDR4 here as this is a budget oriented processor and the current DDR4 prices would blow the balance. Like the predecessor, these new APUs will have four CPUs cores and eight GPU cores in one die.
The biggest highlight is probably that the new APUs will use the latest Excavator cores, these are the 4th generation of the Bulldozer. While we are promised a significant performance boost of 30% from these new cores at their 15W TDPs, most people aren’t too optimistic on the improvements. AMD have surprised us in the past, so I wouldn’t write if off just yet. At the same time experts speculate that the higher wattage CPUs won’t get as good a boost as these entry-level models. The third generation Graphic Core Next (GCN) should also give these tiny APUs a great boost in the graphics department.
The slide above is from earlier this year, but it has all the relevant information about the new architecture collected in one. A thing worth noticing is that the CPU side only has half the L2 cache the Kaveri had. If this is a Die-size issue or it simply wasn’t needed in these models is up to any ones best guess, but this is something that could hurt the performance in some scenarios.
With support for DirectX 12 and HDMI 2.0 these tiny things will be able to handle 4K monitors up to 120Hz. This could make them perfect as HTPC or for product demonstrations. We’ve already seen the Delta Core function mentioned in the slide on the Tonga R9 285 chips. This is a compression technique that has allowed AMD to increase the memory bandwidth by 40%, something that without a doubt will help the APUs who depend on shared DDR3 memory rather than having their own dedicated GDDR5.
It looks like AMD is getting ready for a full-scale launch on the mobile market with the news about Nolan and Amur just a couple of days ago. These two are aimed at 2-in-1 devices and the tablet market and now the Carrizo-L for the entry-level notebooks.
Thank you MyDrivers for providing us with these information
We’ve heard a lot about AMD’s Next-Generation desktop APU platform, codename Carrizo. Last year we heard that the expected arrival date was to be 2015, and up until now we thought this was still on-track especially when details about the A10-8890K APU were released that showed support for a new FM3 socket. New rumours suggest Carrizo may now be delayed for the desktop platform – although the reasons are not entirely clear. AMD will not release Carrizo desktop APUs until 2016 and instead they will refresh the Kaveri product stack to bring out the 5th generation of APUs based on that. Carrizo will still arrive in 2015 in its mobile form according to the source, but there’s no sight of the desktop variant. That’s strange given the roadmap below that leaked from AMD suggesting Carrizo was to be released in 2015.
If the rumour is true Carrizo could now be the 6th generation of AMD APU, not the 5th as that title might get taken by Kaveri refresh. That means like with the Trinity to Richland transition the new Kaveri Refresh parts will be based on virtually identical physical components except they will get speed bumps and some performance and power optimisations. From my perspective this seems like a prudent financial decision being made by AMD: they need another generation of Kaveri APUs to fully profit from all the R&D that went into Kaveri, it also buys them some more time to tweak and tune Carrizo to make that release as competent as possible.
Based on current rumours we think that AMD’s Carrizo will bring a new socket to the table: possibly FM3 or FM3+, but will also retain backwards compatibility with FM2+. It should also introduce a new CPU architecture, possibly based on the Excavator design. It may bring DDR4 support too since DDR4 will be commercially viable from September 2014 when Intel introduces X99, Intel should also introduce DDR4 to its mainstream platforms next year. The other possibility is that Carrizo may opt for stacked-DRAM and DDR3 in order to keep costs down.
A CPU-Z image which appears to detail AMD’s next-generation APU, the A10-8890K, has apparently been leaked. WCCFTech has found the image via Guru3D, but at this point it should be taken as a rumour, since nothing has been officially announced.
The CPU-Z image appears to show that AMD is still keeping the current APU name, having it part of the A10 series, possibly the A-10 Elite. The K suffix at the end shows that the APU in question has an unlocked multiplier and the ‘Elite Hexa-Core’ present in the AMD logo indicates that the APU comes with 6 cores. Also, the 95W TDP and 4.4 Ghz core clock makes it an impressive piece.
The latest AMD APU appears to be based on the FM3 socket, which has been stated to come in 2016, which leads to two possibilities here. The first is that the report from AMD was inconclusive and the APU might come early than reported, or the leaked CPU-Z image does not display an actual Carrizo APU.
The CPU is said to boast Excavator cores, the latest revision of the Bulldozer architecture. This might be the case in the leaked image, having it show a high core clock and number of cores. However, the L2 cache appears to show 3 x 1024 KB, which is said to be very low for a 6-core CPU. AMD might have something up its sleeve with some stacked DRAM lifting the L2 Cache, but nothing is sure at the moment.
Thank you WCCFTech for providing us with this information Images courtesy of WCCFTech
New information has emerged about AMD’s next generation desktop “Carrizo” APUs. We already know so far that Carrizo will be arriving in 2015, it will make use of AMD’s Excavator CPU core architecture, use DDR3 memory, have full HSA support, fit the FM2+ package and work with existing A88X motherboards. The new information suggests AMD’s Carrizo APUs will make use of 28nm APU architecture, that’s covering the CPU and GPU components, but also they will make use of 20nm stacked DRAM.
Why stacked DRAM on-package? Well it certainly does make sense for AMD, even if this information is still in its “rumour” stages. AMD’s APUs are currently bottlenecked by the system memory which is intensively used by the GPU component. Making the move to DDR4 would be expensive so instead stacked DRAM allows DDR3 to be used, but offer better performance than DDR4 because of the shortened and simplified interface pipeline. Stacked DRAM on Carrizo would offer a cost effective way of interfacing with the system memory at higher speeds and lower latencies than DDR4, but also at a lower cost than integrated on-die L3 cache. It would also enable AMD to offer a minimum amount and standard of memory for its HSA concept making it a lot more developer friendly. The stacked DRAM feature is rumoured to be reserved for business-grade and high-end consumer products only: this won’t be equipped on entry-level Carrizo products.
Digitimes reports that AMD has revised its product roadmap for 2014-2015 due to changes in the industry. AMD will proceed with the unveiling of its Volcanic Islands GPU family in late-September 2013. In the first quarter of 2014 they will apparently unveil their fourth generation of APUs codename Kaveri and Kabini. Kabini in socket ST3 for notebooks and FS1B for desktops while Kaveri is still believed to be FM2+. Mass production begins in February 2014 and public announcements in March.
“The Kabini-based APUs will have power consumption of 25W and AMD will release two quad-core models, A4-5350 and A4-5150, as well as a dual-core model E1-2650.”
Kabini’s successor “Beema” is also rescheduled for a launch in the second half of 2014 or first half of 2015 and will also adopt the FS1B socket for AMD’s Heterogeneous Systems Architecture (HSA).
“AMD’s FM1 and AM3 sockets will start phasing out in mid-2013 and the end of 2013, respectively. By the end of 2013, Socket AM3+ processors will account for 30% of AMD’s total processor shipments, while Socket FM2-based processors will account for the remaining 70%.”
In 2015 AMD is expected to release Carrizo based APUs featuring the Excavator architecture with two power consumptions specifications of 45W and 65W, these will replace Kaveri. Nolan will be brought in 2015 to replace Beema.
To clarify potential confusion here is a summary of the changes to the APU roadmap:
High End Desktop APU – Kaveri (Q1 2014) -> Carrizo (2015)
Mobile and Entry Level Desktop APU – Kabini (Q1 2014) -> Beema (H2 2014/H1/2015) – > Nolan (H2 2015)
The future of AMD’s AM3+ processors remains mysteriously unknown. Plans of a successor for the FX AM3+ series is not detailed yet, we may see one, we may not.
Reports suggest that the successor to AMD’s Kaveri platform will be codenamed “Carrizo” as we previously detailed before. Carrizo is the replacement for Kaveri, while Kaveri is the replacement for the current Richland APU. With ASUS having just released their A88XM-A FM2+ motherboard it is expected that both Kaveri and Carrizo APUs will be able to use this socket and chipset and it will last well into 2015.
Kaveri has apparently been delayed to Q2 of 2014 while Carrizo is to be released in 2015. Carrizo will continue to support the A88X and A78 chipsets with TDPs of up to 65W. Apparently Carrizo will support DDR4 memory too.
Kaveri is reported to undergo production in August meaning that mass production can start in December. By that logic we can expect to see Kaveri hit the mainstream in February at the earliest.
According to a report by the Digitimes AMD’s Kaveri APU (Richland’s successor) is due to be replaced by Carizo in 2015. AMD’s Kaveri is due between Q4 of 2012 to Q2 of 2013 and Carizo is due 2015.
“In the first half of 2014, AMD plans to release two more HSA-based APU series: Berlin for server products and Beema, to replace the existing Kabini APU series. In 2015, AMD will release Carizo and Nolan APU series to replace Kaveri and Beema, respectively, both featuring HSA”
Not much else is known about Carizo other than it will offer HSA/HUMA support, that is the ability for the system to have a unified memory architecture for cross-sharing between the GPU and CPU. In terms of the CPU side it is likely that Carizo will be a second version of Steamroller, or a Steamroller refresh, just like how Trinity was Piledriver and Richland was a Piledriver refresh. That said it is also possible Carizo could bring in the new Excavator CPU architecture. In terms of graphics we could see a refresh of the GCN architecture, in the form of GCN 2.0 which seems the most likely option as the HD 9000 series will last well into the end of 2014 and that also uses a variant of GCN (and AMD APUs are known for lagging 1-2 generations behind the desktop AMD GPU platform).
It is still unknown if Carizo will be the “fourth generation” of AMD APUs or just a mere refresh like Richland was. We will have to wait and see for more details. New information is expected next month.