Subject: Processors | May 18, 2017 - 01:01 AM | Tim Verry
Tagged: Zen, Vega, ryzen mobile, ryzen, raven ridge, APU, amd
AMD teased its upcoming Zen-based APUs aimed at mobile devices during its Financial Analyst Day where the company revealed the "Raven Ridge" parts will be aptly known as Ryzen Mobile. The Tech Report managed to acquire a couple slides which confirm some of the broader specifications and reveal how they stack up to AMD's latest Bristol Ridge A-Series APUs – at least as far as AMD's internal testing is concerned (which is to say not independently verified yet so take with a grain of salt).
Ryzen Mobile appears to be the new consumer-facing brand name for what has so far been code named "Raven Ridge". These parts will use a Zen-based CPU, Vega GPU, and integrated chipset. Thanks to the slides, it is now confirmed that the Vega-based graphics processor will be on-die. What has not been confirmed is whether the chipset will be on die or on package and exact specifications on CPU cores counts, GPU Compute Units, cache, memory support, and I/O like PCI-E lanes (you know, all the good stuff! heh). Note that rumors so far point towards Raven Ridge / Ryzen Mobile utilizing a single 4-core (8-thread) CCX, per core L2, 8MB shared L3 cache, and a Vega-based GPU with 1024 cores. HBM2 has also been rumored for awhile but we will have to wait for more leaks and/or an official announcement to know for sure if these Ryzen Mobile parts aimed for the second half of 2017 will have that (hopefully!).
With that said, according to AMD, Ryzen Mobile will offer up to 50% better CPU performance, 40% better GPU performance, and will use up to 50% less power than the previous 7th generation (Excavator-based) A-Series APUs (e.g. FX 9830P and A12-9730P). Those are some pretty bold claims, but still within the realm of possibility. Zen and Vega are both much more efficient architectures and AMD is also benefiting from a smaller process node (TSMC 28nm vs Samsung / GlobalFoundries 14nm FinFET). I do wonder how high the APUs will be able to clock on the CPU side of things with 4 GHz seeming to be the wall for most Zen-based Summit Ridge chips, so most of the CPU performance improvement claims will have to come from architecture changes rather than increases in clockspeeds (the highest clocked A-Series Bristol Ridge ran at up to 3.7 GHz and I would expect Raven Ridge to be around that, maybe the flagship part turbo-ing a bit more). Raven Ridge will benefit from the shared L3 cache and, more importantly, twice as many threads (4 vs 8) and this may be where AMD is primarily getting that 50% more CPU performance number from. On the graphics side of things, it looks like Bristol Ridge with its R7 graphics (GCN 3 (Tonga/Fiji on the Desktop)) had up to 512 cores. Again, taking the rumors into account which say that Raven Ridge will have a 1024 core Vega GPU, this may be where AMD is getting the large performance increase from (the core increase as well as newer architecture). On the other hand, the 40% number could suggest Ryzen Mobile will not have twice the GPU cores. I would guess that 1024 might be possible, but running at lower clocks and that is where the discrepancy is. I will admit I am a bit skeptical about the 1024 (16 CU) number though because that is a huge jump... I guess we will see though!
Further, I am curious if Ryzen Mobile will use HBC (high bandwidth cache) and if HBM2 does turn out to be utilized how that will play into the HBC and whether or not we will finally see the fruits of AMD's HSA labors! I think we will see most systems use DDR4, but certainly some SKUs could use HBM2 and that would definitely open up a lot of performance possibilities on mobile!
There is still a lot that we do not know, but Ryzen Mobile is coming and AMD is making big promises that I hope it delivers on. The company is aiming the new chips at a wide swath of the mobile market from budget laptops and tablets to convertibles and even has their sights set on premium thin and lights. The mobile space is one where AMD has struggled with in getting design wins even when they had good parts for that type of system. They will really need to push and hit Ryzen Mobile out of the park to make inroads into the laptop, tablet, and ultrabook markets!
AMD plans to launch the consumer version of Ryzen Mobile in the second half of this year (presumably with systems featuring the new APUs out in time for the holidays if not for the back to school end of summer rush). The commercial SKUs (which I think refers to the Ryzen equivalent of AMD Pro series APUs.Update: Mobile Ryzen Pro) will follow in the first half of 2018.
What are your thoughts on Ryzen Mobile and the alleged performance and power characteristics? Do you think the rumors are looking more or less correct?
- Zen and the Art of CPU Design
- AMD Launching Ryzen 5 Six Core Processors Soon (Q2 2017)
- AMD Vega GPU Architecture Preview: Redesigned Memory Architecture
- The AMD Ryzen 7 1800X Review: Now and Zen
- More Ryzen coverage!
Subject: Processors | May 17, 2017 - 04:05 AM | Scott Michaud
Tagged: amd, EPYC, 32 core, 64 thread, Intel, Broadwell-E, xeon
AMD has formally announced their EPYC CPUs. While Sebastian covered the product specifications, AMD has also released performance claims against a pair of Intel’s Broadwell-E Xeons. While Intel’s E5-2650 v4 processors have an MSRP of around $1170 USD, each, we don’t know how that price will compare to AMD’s offering. At first glance, pitting thirty two cores against two twelve-core chips seems a bit unfair, although it could end up being a very fair comparison if the prices align.
Image Credit: Patrick Moorhead
Patrick Moorhead, who was at the event, tweeted out photos of a benchmark where Ubuntu was compiled over GCC. It looks like EPYC completed in just 33.7s while the Broadwell-E chip took 37.2s (making AMD’s part ~9.5% faster). While this, again, stems from having a third more cores, this depends on how much AMD is going to charge you for them, versus Intel’s current pricing structure.
Image Credit: Patrick Moorhead
This one chip also has 128 PCIe lanes, rather than Intel’s 80 total lanes spread across two chips.
Subject: Processors | May 16, 2017 - 07:22 PM | Sebastian Peak
Tagged: Zen, Threadripper, ryzen, processor, HEDT, cpu, amd
AMD revealed their entry into high-end desktop (HEDT) with the upcoming Ryzen "Threadripper" CPUs, which will feature up to 16 cores and 32 threads.
Little information was revealed along with the announcement, other than to announce availablility as "summer 2017", though rumors and leaks surrounding Threadripper have been seen on the internet (naturally) leading up to today's announcement, including this one from Wccftech. Not only will Threadripper (allegedly) offer quad-channel memory support and 44 PCI Express lanes, but they are also rumored to be released in a massive 4094-pin package (same as "Naples" aka EPYC) that most assuredly will not fit into the AM4 socket.
Image credit: Wccftech
These Threadripper CPUs follow the lead of Intel's HEDT parts on X99, which are essentially re-appropriated Xeons with higher clock speeds and some feature differences such as a lack of ECC memory support. It remains to be seen what exactly will separate the enthusiast AMD platform from the EPYC datacenter platform, though the rumored base clock speeds are much higher with Threadripper.
Subject: Processors | May 16, 2017 - 06:49 PM | Sebastian Peak
Tagged: Zen, server, ryzen, processor, EPYC, datacenter, cpu, amd, 64 thread, 32 core
AMD has announced their new datacenter CPU built on the Zen architecture, which the company is calling EPYC. And epic they are, as these server processors will be offered with up to 32 cores and 64 threads, 8 memory channels, and 128 PCI Express lanes per CPU.
Some of the details about the upcoming "Naples" server processors (now EPYC) were revealed by AMD back in March, when the upcoming server chips were previewed:
- A highly scalable, 32-core System on Chip (SoC) design, with support for two high-performance threads per core
- Industry-leading memory bandwidth, with 8-channels of memory per "Naples" device. In a 2-socket server, support for up to 32 DIMMS of DDR4 on 16 memory channels, delivering up to 4 terabytes of total memory capacity.
- The processor is a complete SoC with fully integrated, high-speed I/O supporting 128 lanes of PCIe, negating the need for a separate chip-set
- A highly-optimized cache structure for high-performance, energy efficient compute
- AMD Infinity Fabric coherent interconnect for two "Naples" CPUs in a 2-socket system
- Dedicated security hardware
Compared to Ryzen (or should it be RYZEN?), EPYC offers a huge jump in core count and available performance - though AMD's other CPU announcement (Threadripper) bridges the gap between the desktop and datacenter offerings with an HEDT product. This also serves to bring AMD's CPU offerings to parity with the Intel product stack with desktop/high performance desktop/server CPUs.
EPYC is a large processor. (Image credit: The Tech Report)
While specifications were not offered, there have been leaks (of course) to help fill in the blanks. Wccftech offers these specs for EPYC (on the left):
(Image credit: Wccftech)
We await further information from AMD about the EPYC launch.
Application Profiling Tells the Story
It should come as no surprise to anyone that has been paying attention the last two months that the latest AMD Ryzen processors and architecture are getting a lot of attention. Ryzen 7 launched with a $499 part that bested the Intel $1000 CPU at heavily threaded applications and Ryzen 5 launched with great value as well, positioning a 6-core/12-thread CPU against quad-core parts from the competition. But part of the story that permeated through both the Ryzen 7 and the Ryzen 5 processor launches was the situation surrounding gaming performance, in particular 1080p gaming, and the surprising delta that we see in some games.
Our team has done quite a bit of research and testing on this topic. This included a detailed look at the first asserted reason for the performance gap, the Windows 10 scheduler. Our summary there was that the scheduler was working as expected and that minimal difference was seen when moving between different power modes. We also talked directly with AMD to find out its then current stance on the results, backing up our claims on the scheduler and presented a better outlook for gaming going forward. When AMD wanted to test a new custom Windows 10 power profile to help improve performance in some cases, we took part in that too. In late March we saw the first gaming performance update occur courtesy of Ashes of the Singularity: Escalation where an engine update to utilize more threads resulted in as much as 31% average frame increase.
As a part of that dissection of the Windows 10 scheduler story, we also discovered interesting data about the CCX construction and how the two modules on the 1800X communicated. The result was significantly longer thread to thread latencies than we had seen in any platform before and it was because of the fabric implementation that AMD integrated with the Zen architecture.
This has led me down another hole recently, wondering if we could further compartmentalize the gaming performance of the Ryzen processors using memory latency. As I showed in my Ryzen 5 review, memory frequency and throughput directly correlates to gaming performance improvements, in the order of 14% in some cases. But what about looking solely at memory latency alone?
Subject: Processors | May 9, 2017 - 03:26 PM | Jeremy Hellstrom
Tagged: ryzen, amd, 1500X, 1600X, ryzen 5
The pricing of AMD's Ryzen 5 line spans from $170 to $250, similar to Intel's Core i5 line and may wwll tempt those a generation or two out of date to consider an upgrade. In order to demonstrate differences in CPU performance Ars Technica tested both Intel and AMD processors paired with a GTX 1080 Ti. By doing so at lower resolutions which the card can more than handle they expose differences in the performance of the two architectures, which seem to follow AMD's offerings into higher resolutions albeit with a smaller performance delta. Check out the wide gamut of tests that were performed to see which architecture makes more sense for your usage, especially if you do more than just gaming and surfing.
"The Ryzen 5 range is made up of four chips. At the top is the £240/$250 Ryzen 5 1600X, a 95W six-core chip that boasts simultaneous multithreading (SMT, the equivalent of hyper-threading), 16MB of L3 cache, and a 3.6GHz base clock."
Here are some more Processor articles from around the web:
- The AMD Ryzen 7 1800X Octa-Core @ TechARP
- The Complete AMD Ryzen 7 Tech Report @ TechARP
- Pentium G4560 CPU @ Hardware Secrets
Subject: Processors | May 9, 2017 - 03:13 PM | Ryan Shrout
Tagged: ryzen, amd, 1700X
A little birdie sent me a note this afternoon that the AMD Ryzen 7 1700X processor was selling on Amazon.com for just $333! Considering the launch price of that CPU was $399 just two months ago, a $60-70 discount makes this platform all the more compelling for consumers looking to build a new PC. Coupling that with the overclocking performance we saw from our Ryzen 1700 sample, you should be able to meet or exceed expectations with the 1700X model.
This link led me down a bit of a rabbit hole as I wanted to see where a solid build would stand using that processor and a focus on budget. Now, keep in mind that this was put together rather hastily this afternoon, but here's what I came up with.
|Ryzen 7 1700X Build|
|Processor||AMD Ryzen 7 1700X - $333|
|Cooler||Thermaltake Contac Silent - $24|
|Motherboard||ASUS Prime B350-Plus - $99|
|Memory||G.Skill Ripjaws 16GB DDR4-3000 - $118|
|Graphics Card||EVGA GeForce GTX 1050 Ti 4GB - $149|
|Storage||Samsung 850 EVO 250GB - $107|
|Case||Corsair 200R ATX Mid Tower Case - $56|
|Power Supply||Corsair CX 500 watt - $59|
|Total Price||$945 - Amazon.com Full Cart|
For the base of the system, you can pick up the Ryzen 7 1700X processor for $333, a great B350-based motherboard from ASUS for $99 and 16GB of DDR4 memory running at 3000 MHz for just $118. Getting that memory at higher clock speeds is important for optimal Ryzen performance - hunt around to find the best deal! That's just $550 for the heart of a system that could power anything from the GTX 1050 Ti I included above to the GTX 1080 Ti if you are pushing the limits of graphics performance.
If you try to stay within a reasonable budget, as I did above, you can build a from-scratch machine for under $1000 with some impressive specifications and capabilities. The GTX 1050 Ti will get you peak 1080p gaming capability while the 8-cores and 16-threads of the Ryzen 7 1700X will improve any workstation-class or multimedia workloads.
Separately, but interestingly, the gang at 3DCenter.org posted the results of a survey taken about the Ryzen 5 processor launch, measuring the readers reactions to the release. In it, 83.9% of the audience looked upon the Ryzen 5 favorably, 9.4% as average and 6.7% negatively. If you compare that to the Ryzen 7 launch (74.6% favorable, 17.5% average, 7.9% negative) it seems that Ryzen 5 was better received than its big brother. But if you look back to October 2011 when the same survey was run about AMD Bulldozer, only 6.8% saw the CPU launch as favorable (!!). The last CPU launch that received nearly as positive a reaction as Ryzen 5/7 was the Sandy Bridge CPU back in January of 2011.
Obviously this survey isn't a predictor of success or failure exactly, but it does point to an audience that is incredibly receptive to the new AMD processors. My own experience tells me that these numbers are fairly accurate to the mood about Ryzen, even after the 1080p gaming fiasco that circulates to this day. Interest and reaction are great for a company that needs to make in-roads in the market, but converting that consumer interest into purchases is the key for AMD going forward.
Subject: Processors, Mobile | May 9, 2017 - 01:55 PM | Ryan Shrout
Tagged: spectra, snapdragon mobile platform, snapdragon, qualcomm, Kryo, isp, hexagon, dsp, adreno, 660, 630
Today Qualcomm took the covers off of an update to the Snapdragon 600 family of processors, now known as mobile platforms. The Snapdragon 660 and 630 Mobile Platforms are important products in the company’s portfolio as they address a larger segment of the consumer market than the premium-tier Snapdragon 800 while still offering performance and feature sets above the budget segments of the 400s. The Snapdragon 820 and 835 traditionally get all of the attention from media, the 600-series is at the heart of popular devices like the Sony Xperia X, Asus Zenfone 3 Ultra, HTC 10 Lifestyle and over 1000 more designs.
The biggest changes to both new platforms come in the form of LTE connectivity and GPU performance. In a bid to bring previously unseen capabilities to the 600-series of solutions, Qualcomm has taken the Snapdragon X12 LTE modem that shipped with the Snapdragon 820/821 SoC and integrated it on both the 660 and 630. This creates mainstream mobile platforms that can run Cat 12/13 modems and speeds as high as 600 Mbps downstream (3x carrier aggregation) and 150 Mbps upstream (2x carrier aggregation).
That is a significant move and should result in a massive amount of high speed devices saturating the market (and carriers’ networks) starting later this year. Along with that higher performance comes the same X12 feature set that we saw with Snapdragon 820/821 including adaptive antenna tuning capability (TruSignal) and dynamic signal quality adjustments for power efficiency optimizations.
The GPU performance of both the Snapdragon 660 and 630 get a boost over the previous competitors (653 and 626 respectively) though they do so with different Adreno implementations. The SD 660 uses the Adreno 512 GPU that offers up to 30% better performance compared to the Adreno 510 used on the SD 650 series. While we don’t have details yet on where that advantage comes from (clocks or core improvements), I have a feeling that much of it comes from improved frequencies. The Snapdragon 630 uses the Adreno 508 GPU, compared to the 506 from the SD 626 processor, and also claims to have a 30% performance advantage over the previous generation.
Subject: Processors | April 27, 2017 - 06:24 PM | Jeremy Hellstrom
Tagged: rockit, risky business, overclocking, kaby lake, delidding, core i7 7700k, aqua computer
Delidding a Kaby Lake processor such as the i7-7700k does not offer the same overclocking advantages as with previous generations when replacing the TIM gave you more headroom. Instead of being able to push your CPU past 5GHz, popping the lid off of a Kaby Lake reduces operating temperatures and likely extends the life of the processor ... or immediately ends it. If you don't have a 3D printer handy to make your own delidder, then take a peek at this review from TechPowerUp. They try out two delidding tools, one from Aqua Computer and one from Rockit which Morry has used; do be aware that any CPUs killed as a result of reading their review is the responsibility of the one who delidded.
"Intel mainstream CPUs have had a bottleneck in cooling due to poor heat transfer from the CPU die to the integrated heat spreader. Thanks to new de-lidding friendly tools released recently, it is now easier than ever before to handle this yourself and get a cooler running CPU. We examine two such solutions from Rockit Cool and Aqua Computer today, both of which promise fool-proof de-lidding and re-lidding"
Here are some more Processor articles from around the web:
- Ryzen For The Masses: A Look At AMD’s Ryzen 5 1600X & 1500X Processors @ Techgage
- AMD Ryzen R5 1600X @ eTeknix
- AMD Ryzen R5 1500X @ Kitguru
- An In-depth Look At AMD’s Ryzen 7 1800X, 1700X & 1700 Processors @ Techgage
Introduction and Technical Specifications
Courtesy of EKWB
EK's Supremacy line of CPU waterblocks are well known for their performance and style. Their latest version in this block line, the Supremacy MX, advances their design in the hopes of getting more optimized performance out of a less costly version of their award winning block series. The base Supremacy MX CPU waterblock is a copper and plexi construction using the same jet-impingement and micro-channel design as that used in their previous block versions. The block comes fully assembled from the factory with a single CPU mounting bracket type (in this case, the Intel version). Note that additional CPU mounting kits are available for purchase. With an MSRP of $54.99, the Supremacy MX waterblock offers a compelling purchase in light of its performance potential.
Courtesy of EKWB
Courtesy of EKWB
The block is assembled with hex-head screws going through the copper base plate with rubber grommets ensuring the integrity of the block internals. The top aluminum cover plate is held to the plexi top using short hex-head screws that thread directly into the plexi top plate. The center inlet feeds the micro-channels embedded in the copper base plate through the jet-impingement assembly. The mounting bracket sits in between the top plexi plate and the copper base plate, making any an interesting upgrade if you want to switch out the CPU mount plate to use the block on a different CPU family (like going from Intel to AMD Ryzen for example). The aluminum top plate gives the block a sleek appearance and acts to redirect illumination from the side mounted LEDs (if you choose to use LEDs with the block that is).