Subject: Processors | February 19, 2019 - 02:57 PM | Tim Verry
Tagged: Zen 2, x570, X500, Ryzen 3000, navi, matisse, amd, 7nm
Spotted by HardOCP, Paul from Red Gaming Tech recently shared leaked information from a source with a reputation of being reliable (from past leaks about 7nm GPUs) who claims that AMD will be announcing a plethora of products at Computex in June to setup for the launch of Zen 2-based 7nm "Matisse" Ryzen 3000 desktop processors, X500 series chipset-based motherboards, and 7nm Navi-based consumer gaming graphics cards on July 7th (The 7th for 7nm I guess).
Image via AnandTech
As a refresher, Zen 2 is the next major architectural jump for AMD while also pushing a new smaller process node. AMD has not yet revealed all the details about Zen 2 especially about consumer chips, but the new microarchitecture is said to feature tweaks to the front end that along with clockspeed bumps from the TSMC 7nm process will allow them to realize notable IPC and single threaded performance gains. When talking about EPYC 2 "Rome" server processors (Zen 2 based) AMD hinted at changes to branch prediction and pre-fetching as well as increased cache sizes and larger FPUs (256-bit), for example. The move to 7nm allegedly allows AMD to hit similar power envelopes to Zen+ (12nm) Ryzen 2000 series processors while hitting much higher clockspeeds at up to 5.1 GHz boost on their top-end chip. While mobile chips may strike a finer balance between power usage and performance with the move to 7nm, on the desktop AMD is spending nearly all the power savings on performance (which makes sense). Note that it is still not officially official that AMD is using a scaled down EPYC setup with more than one 7nm (TSMC) CPU die and a separate IO die (14nm Global Foundries), [they only teased a chip at CES with an IO die and a single CPU die] but I am of the opinion that that particular rumor makes more sense than otherwise so am inclined to believe this is the case.
Ryzen 3000 series processors feature an IO chiplet along with what is rumored to be up to two CPU chiplets (image credit: Tom's Hardware).
From previous leaks, Ryzen 3000 is said to cover all the bases from six core Ryzen 3 3300 series chips to midrange eight core Ryzen 5 and on up to 12 and 16 core Ryzen 9 CPUs that move beyond a single CPU die to two 7nm CPU dies that feature eight cores each. In fact, the top end Ryzen 9 3850X is supposedly a 16 core (32 thread) monster of a desktop chip that has a base frequency of 4.3 GHz and can boost up to 5.1 GHz with a 135W TDP (which when overclocked will likely draw dramatically more like we've seen with both AMD and Intel's top end consumer chips) and price tag of around $520 (400 pounds). The Ryzen 7 3700 and 3700X are 12 core (24 thread) models with TDPs of 95W and 105W respectively with the non-x SKU clocked at 3.8 to 4.6 GHz and the 3700X clocked at 4.2 GHz base and 5 GHz boost. The Ryzen 5 3600 and 3600X are the top end single CPU die models (though a 2x single CCX per die chips might be a reality depending on yields) at eight cores and 16 threads. The Ryzen 3 3300 series parts represent the low end which is now interestingly six cores (oh how times have changed!). Perhaps most interesting of the leaked chips are the Ryzen 5 3600G (~$207) and the Ryzen 3 3300G (~$130) though which feature Navi 12 integrated graphics (presumably these processors combine one 7nm CPU die, one 7nm GPU die, and one 14nm IO die) with 15 and 20 CUs respectively.
As for motherboards, in general the new chips will use the AM4 socket and will be compatible with older 300 and 400 series motherboards with a BIOS update though the top end chips may well necessitate a new X570 or other X500 series motherboard with better power delivery especially for enthusiasts planning to attempt stable overclocks.
Unfortunately, on Navi details are still a bit scarce but the new architecture should bring performance enhancements even beyond Radeon VII (Vega on 7nm). Allegedly due to issues with TSMC, Red Gaming Tech's source believes that Navi might be delayed or pushed back beyond the planned mid-summer release date, but we will have to wait and see. As TSMC ramps up its partial EUV enhanced 7nm node it may free up needed production line space of the current 7nm node for AMD (to fight with others over heh) to meet its intended deadline but we will just have to wait and see!
Take these rumors with a grain of salt as usual but it certainly sounds like it is hoing to be an exciting summer for PC hardware! Hopefully more details about Ryzen 3000 and Navi emerge before then though as that's quite a while yet to wait. Of course, Zen 2 APUs are not coming until at least next year and AMD is still not talking Zen 2 Threadripper which may not see release until the fall at the very earliest. I am very interested to see how AMDs chiplet based design fares and how well they are able to scale it across their product stack(s) as well as what Intel's response will be as it presses on with a fine tuned 14nm++ and a less ambitious 10nm node.
- AMD Shows Off Zen 2-Based EPYC "Rome" Server Processor
- Podcast #521 - Zen 2, 7nm Vega, SSD Vulnerabilities, and more!
- Nein, Zeneration 3 is best
Subject: Graphics Cards | February 7, 2019 - 03:30 PM | Jeremy Hellstrom
Tagged: VRAM, video card, Vega 20, Vega, radeon vii, radeon, pcie, opencl, HBM2, graphics card, gaming, compute, amd, 7nm, 16GB
While enjoying the pictures and tests Sebastian ran on the new AMD Radeon VII, was there a game that we missed that is near and dear to your heart? Then perhaps one of these reviews below will solve that, the list even includes Linux performance for those on that side of the silicon. For instance, over at The Tech Report you can check out Monster Hunter: World, Forza Horizon 4 and the impressive results that the new 7nm card offers in Battlefield V.
"AMD's Radeon VII is the first gaming graphics card powered by a 7 nm GPU: Vega 20. This hopped-up Vega chip comes linked up with 16 GB of HBM2 RAM good for 1 TB/s of memory bandwidth. We put this potent combination to the test to see if it can beat out Nvidia's GeForce RTX 2080."
Here are some more Graphics Card articles from around the web:
- AMD Radeon VII @ Guru of 3D
- AMD Radeon VII 16GB Video Card Review @ Legit Reviews
- AMD Radeon VII: A 7nm-long step in the right direction, but is that enough? @ Ars Technica
- AMD Radeon VII 1440p, 4K & Ultrawide Gaming Performance @ Techgage
- AMD Radeon VII Review: RTX Killer or Flop? @ Techspot
- AMD Radeon VII 16 GB @ TechPowerUp
- AMD Radeon VII @ Kitguru
- AMD Radeon VII Linux Benchmarks - Powerful Open-Source Graphics For Compute & Gaming @ Phoronix
Overview and Specifications
After a month-long wait following its announcement during the AMD keynote at CES, the Radeon VII is finally here. By now you probably know that this is the world’s first 7nm gaming GPU, and it is launching today at a price equal to NVIDIA’s GeForce RTX 2080 at $699.
The AMD Radeon VII in action on the test bench
More than a gaming card, the Radeon VII is being positioned as a card for content creators as well by AMD, with its 16GB of fast HBM2 memory and enhanced compute capabilities complimenting what should be significantly improved gaming performance compared to the RX Vega 64.
Vega at 7nm
At the heart of the Radeon VII is the Vega 20 GPU, introduced with the Radeon Instinct MI60 and MI50 compute cards for the professional market back in November. The move to 7nm brings a reduction in die size from 495 mm2 with Vega 10 to 331 mm2 with Vega 20, but this new GPU is more than a die shrink with the most notable improvement by way of memory throughput, as this is significantly higher with Vega 20.
Double the HBM2, more than double the bandwidth
While effective memory speeds have been improved only slightly from 1.89 Gbps to 2.0 Gbps, far more impactful is the addition of two 4GB HBM2 stacks which not only increase the total memory to 16GB, but bring with them two additional memory controllers which double the interface width from 2048-bit to 4096-bit. This provides a whopping 1TB (1024 GB/s) of memory bandwidth, up from 483.8 GB/s with the RX Vega 64.
Subject: Graphics Cards | January 12, 2019 - 08:17 AM | Jim Tanous
Tagged: vega 64, Vega, RX VEGA 64, radeon vii, gpu, benchmarks, amd, 7nm
After announcing the Radeon VII this week at CES, AMD has quietly released its own internal benchmarks showing how the upcoming card potentially compares to the Radeon RX Vega 64, AMD's current flagship desktop GPU released in August 2017.
The internal benchmarks, compiled by AMD Performance Labs earlier this month, were released as a footnote in AMD's official Radeon VII press release and first noticed by HardOCP. AMD tested 25 games and 4 media creation applications, with the Radeon VII averaging around a 29 percent improvement in games and 36 percent improvement in professional apps.
AMD's test platform for its gaming Radeon VII benchmarks was an Intel Core i7-7700K with 16GB of DDR4 memory clocked at 3000MHz running Windows 10 with AMD Driver version 18.50. CPU frequencies and exact Windows 10 version were not disclosed. AMD states that all games were run at "4K max settings" with reported frame rate results based on the average of three separate runs each.
For games, the Radeon VII benchmarks show a wide performance delta compared to RX Vega 64, from as little as 7.5 percent in Hitman 2 to as much as 68.4 percent for Fallout 76. Below is a chart created by PC Perspective from AMD's data of the frame rate results from all 25 games.
In terms of media creation applications, AMD changed its testing platform to the Ryzen 7 2700X, also paired with 16GB of DDR4 at 3000MHz. Again, exact processor frequencies and other details were not disclosed. The results reveal between a 27% and 62% improvement:
It is important to reiterate that the data presented in the above charts is from AMD's own internal testing, and should therefore be viewed skeptically until third party Radeon VII benchmarks are available. However, these benchmarks do provide an interesting first look at potential Radeon VII performance compared to its predecessor.
Radeon VII is scheduled to launch February 7, 2019 with an MSRP of $699. In addition to the reference design showcased at CES, AMD has confirmed that third party Radeon VII boards will be available from the company's GPU partners.
Subject: Graphics Cards | January 11, 2019 - 04:35 PM | Sebastian Peak
Tagged: video cards, Vega VII, Vega, Refresh, radeon, Mark Papermaster, graphics, gpus, cto, amd, 7nm
AMD CTO Mark Papermaster spoke with The Street in a video interview published yesterday, where he made it clear that we can indeed expect a new Radeon lineup this year. “It’s like what we do every year,” he said, “we’ll round out the whole roadmap”.
Part of this refresh has already been announced, of course, as Papermaster noted, “we’re really excited to start on the high end” (speaking about the Radeon VII) and he concluded with the promise that “you’ll see the announcements over the course of the year as we refresh across our Radeon roadmap”. It was not mentioned if the refreshed lineup will include 7 nm parts derived from the Vega VII shown at CES, but it seems reasonable to assume that we haven’t seen the last of Vega 2 in 2019.
Subject: General Tech | January 11, 2019 - 01:14 PM | Jeremy Hellstrom
Tagged: amd, 7nm, CoWoS, TSMC, SPIL, TFME
DigiTimes today is sharing some information about just where AMD's 7nm chips will be processed and there seems to be a name missing. TSMC, SPIL and TFME will all be producing specific products but there is no mention of GLOBALFOUNDRIES in the news post.
TSMC will handle the bulk of the EPYC and HPC versions of Vega production with their chip-on-wafer-on-substrate, as one might expect; SPIL and TFME will handle desktop Ryzen and GPUs. One hopes that by diversifying their production sources we can avoid shortages from one line effecting the entire market as we have seen in the past.
"TSMC is also among the backend partners of AMD for its new 7nm computing and graphics products, according to industry sources. Siliconware Precision Industries (SPIL) under Taiwan's ASE Technology Holding, and China-based Tongfu Microelectronics (TFME) are other backend service providers for the chips, the sources continued."
Here is some more Tech News from around the web:
- Windows 10 Insiders sent on quest deep into Registry to fetch goblet of Reserved Storage @ The Register
- After broken promise, AT&T says it’ll stop selling phone location data @ Ars Technica
- Developer Bungie Splits With Publisher Activision, Will Keep World Shooter Series Destiny @ Slashdot
- Reddit locks out users with poor password hygiene after spotting 'unusual activity' @ The Register
- PC shipments drop again, just as things were starting to look up @ The Inquirer
- Sex toy wins CES robotics award, then has it taken away in ridiculous moral panic @ The Inquirer
- Hackaday Podcast Ep1 – Seriously, We Know What We’re Doing
Subject: Graphics Cards | December 10, 2018 - 03:28 PM | Sebastian Peak
Tagged: Vega, trademark, rumor, report, radeon, graphics, gpu, amd, 7nm
The logo, with the familiar "V" joined by a couple of new stripes on the right side, could mean a couple of things; with a possible reference to Vega II (2), or perhaps the VII suggests the Roman numeral 7 for 7nm, instead? VideoCardz.com thinks the latter may be the case:
"AMD has registered a new trademark just 2 weeks ago. Despite many rumors floating around about Navi architecture and its possible early reveal or announcement in January, it seems that AMD is not yet done with Vega. The Radeon Vega logo, which features the distinctive V lettering, has now received 2 stripes, to indicate the 7nm die shrink."
Whatever the case may be it's interesting to consider the possibility of a 7nm Vega GPU before we see Navi. We really don't know, though it does seem a bit presumptuous to consider a new product as early as CES, as Tech Radar speculates:
"We know full well that the next generation of AMD graphics will be built upon a 7nm architecture going by the roadmaps the company released at CES 2018. At the same time, it seems to all sync up with AMD's plans to announce new 7nm GPUs at CES 2019, so it almost seems certain that we’ll see Vega II graphics cards soon."
The prospect of new graphics cards is always tantalizing, but we'll need more than a logo before things really get interesting.
Subject: General Tech | November 8, 2018 - 01:54 PM | Ken Addison
Tagged: Zen 2, xeon, Vega, rome, radeon instinct, podcast, MI60, Intel, EPYC, cxl-ap, chiplet, cascade lake, amd, 7nm
PC Perspective Podcast #521 - 11/08/18
Join us this week for discussion on AMD's new Zen 2 architecture, 7nm Vega GPUs, SSD encryption vulnerabilities, and more!
The URL for the podcast is: http://pcper.com/podcast - Share with your friends!
- iTunes - Subscribe to the podcast directly through the iTunes Store (audio only)
- Google Play - Subscribe to our audio podcast directly through Google Play!
- RSS - Subscribe through your regular RSS reader (audio only)
- MP3 - Direct download link to the MP3 file
Hosts: Jim Tanous, Jeremy Hellstrom, Josh Walrath, Allyn Malventano, Ken Addison, and Sebastian Peak
Peanut Gallery: Alex Lustenberg
Program length: 1:42:27
Podcast topics of discussion:
Week in Review:
Thanks to Casper for supporting our podcast! Save $50 on select mattresses at http://www.casper.com/pcper code pcper
News items of interest:
Picks of the Week:
Subject: Processors | November 7, 2018 - 11:00 PM | Tim Verry
Tagged: Zen 2, rome, PCI-e 4, Infinity Fabric, EPYC, ddr4, amd, 7nm
In addition to AMD's reveal of 7nm GPUs used in its Radeon Instinct MI60 and MI50 graphics cards (aimed at machine learning and other HPC acceleration), the company teased a few morsels of information on its 7nm CPUs. Specifically, AMD teased attendees of its New Horizon event with information on its 7nm "Rome" EPYC processors based on the new Zen 2 architecture.
Tom's Hardware spotted the upcoming Epyc processor at AMD's New Horizon event.
The codenamed "Rome" EPYC processors will utilize a MCM design like its EPYC and Threadripper predecessors, but increases the number of CPU dies from four to eight (with each chiplet containing eight cores with two CCXs) and adds a new 14nm I/O die that sits in the center of processor that consolidates memory and I/O channels to help even-out the latency among all the cores of the various dies. This new approach allows each chip to directly access up to eight channels of DDR4 memory (up to 4TB) and will no longer have to send requests to neighboring dies connected to memory which was the case with, for example, Threadripper 2. The I/O die is speculated by TechPowerUp to also be responsible for other I/O duties such as PCI-E 4.0 and the PCH communication duties previously integrated into each die.
"Rome" EPYC processors with up to 64 cores (128 threads) are expected to launch next year with AMD already sampling processors to its biggest enterprise clients. The new Zen 2-based processors should work with existing Naples and future Milan server platforms. EPYC will feature from four to up to eight 7nm Zen 2 dies connected via Infinity Fabric to a 14nm I/O die.
AMD CEO Lisa Su holding up "Rome" EPYC CPU during press conference earlier this year.
The new 7nm Zen 2 CPU dies are much smaller than the dies of previous generation parts (even 12nm Zen+). AMD has not provided full details on the changes it has made with the new Zen 2 architecutre, but it has apparently heavily tweaked the front end operations (branch prediction, pre-fetching) and increased cache sizes as well as doubling the size of the FPUs to 256-bit. The architectural improvements alogn with the die shrink should allow AMD to show off some respectable IPC improvements and I am interested to see details and how Zen 2 will shake out.
Subject: General Tech | November 6, 2018 - 03:42 PM | Jeremy Hellstrom
Tagged: AMD Radeon Instinct, MI60, MI50, 7nm, ROCm 2.0, HPC, amd
If you haven't been watching AMD's launch of the 7nm Vega based MI60 and MI50 then you can catch up right here.
You won't be gaming with these beasts, but for those working on deep learning, HPC, cloud computing or rendering apps you might want to take a deeper look. The new PCIe 4.0 cards use HBM2 ECC memory and Infinity Fabric interconnects, offering up to 1 TB/s of memory bandwidth.
The MI60 features 32GB of HBM2 with 64 Compute Units containing 4096 Stream Processors which translates into 59 TOPS INT8, up to 29.5 TFLOPS FP16, 14.7 TFLOPS FP32 and 7.4 TFLOPS FP64. AMD claims is currently the fastest double precision PCIe card on the market, with the 16GB Tesla V100 offering 7 TFLOPS of FP64 performance.
The MI50 is a little less powerful though with 16GB of HBM2, 53.6 TFLOPS of INT8, up to 26.8 TFLOPS FP16, 13.4 TFLOPS FP32 and 6.7 TFLOPS FP64 it is no slouch.
With two Infinity Fabric links per GPU, they can deliver up to 200 GB/s of peer-to-peer bandwidth and you can configure up to four GPUs in a hive ring configuration, made of two hives in eight GPU servers with the help of the new ROCm 2.0 software.
Expect to see AMD in more HPC servers starting at the beginning of the new year, when they start shipping.