Category: Hardware

Intel unveils its first chips built for AI in the cloud

Intel is no stranger to AI-oriented chips, but now it’s turning its attention to those chips that might be thousands of miles away. The tech firm has introduced two new Nervana Neural Network Processors, the NNP-T1000 (below) and NNP-I1000 (above), that are Intel’s first ASICs designed explicitly for AI in the cloud. The NNT-T chip is meant for training AIs in a ‘balanced’ design that can scale from small computer clusters through to supercomputers, while the NNP-I model handles “intense” inference tasks.

The chipmaker also unveiled a next-gen Movidius Vision Processing Unit whose updated computer vision architecture promises over 10 times the inference performance while reportedly managing efficiency six times better than rivals. Those claims have yet to pan out in the real world, but it’s safe to presume that anyone relying on Intel tech for visual AI work will want to give this a look.

You’ll have to be patient for the Movidius chip when it won’t ship until sometime in the first half of 2020. This could nonetheless represent a big leap for AI performance, at least among companies that aren’t relying on rivals like NVIDIA. Intel warned that bleeding-edge uses of AI could require performance to double every 3.5 months — that’s not going to happen if companies simply rely on conventional CPUs. And when internet giants like Facebook and Baidu lean heavily on Intel for AI, you might see practical benefits like faster site loads or more advanced AI features.

AMD Ryzen 9 APUs could present a new challenge for Intel

AMD has been doing everything right and is looking to parlay its recent victories into sustained success for the long haul. According to hardware leaker @Komachi, the chipmaker could do just that through the use of integrated graphics.

As TechRadar highlights, a list of processors recently published by the hardware leaker highlights four Ryzen 9 parts – two 45W units and two 15W “Pro” models – sporting a B12 designation. This, according to the leaker, indicates the presence of 12 graphical compute units.

Ryzen 9 APUs, as the publication points out, could benefit from heightened clock speeds and increased efficiency afforded by their 7-nanometer design. This, in turn, would presumably benefit graphics compute units as well. It’d be an interesting alternative to discrete graphics solutions and would of course further press Intel in the portable laptop space.

AMD has had a phenomenal 2019, recently reporting its highest quarterly revenue since 2015. The company still has a lot of ground to make up but has been steering the ship in the right direction with increased market share ever since Q2 2017.

As always, it’s worth reiterating that the post from @Komachi neither confirms nor denies anything as nothing is truly official until official word comes from the horse’s mouth. With CES 2020 less than two months away, however, the timing is certainly right for a hardware announcement.

NVIDIA Introducing Ampere in March 2020 & Releasing RTX 3080 in June, Analyst Indicates

NVIDIA is allegedly preparing to introduce its next generation family of 7nm graphics cards based on the Ampere architecture next year at GTC in late March and is reportedly planning to release the RTX 3080 around Computex 2020 in June.

There has been a lot of talk in the techsphere lately about NVIDIA’s next generation Ampere graphics cards, which was initially sparked by the company allegedly communicating some initial marketing and go-to-market strategy for Ampere. 

NVIDIA Allegedly Introducing Next Generation 7nm Ampere GPUs in March 2020 at GTC

According to several rumors floating around, NVIDIA CEO & co-founder Jensen Huang is said to be not in too much of a hurry to bring Ampere to market just yet, largely due to how well positioned Turing already is in the market against the competition from AMD. Especially in the high-end segment where the RTX 2080 and RTX 2080 Ti sit uncontested.

However, with AMD preparing its own alleged “NVIDIA Killer” high-end ray tracing graphics card for next year, team green is getting ready to preempt the red team with Ampere in the first half of 2020.

According to Raymond James analyst Chris Caso via HKEPC, NVIDIA has allegedly pushed Ampere back a few months, which is currently in the beginning of its ramp. Ampere is now set mainly for an announcement at GTC next year in late March. According to Caso NVIDIA plans to kick off Ampere’s introduction with a new datacenter focused AI product first , i.e. an Ampere based TESLA accelerator at GTC.

NVIDIA Allegedly Launching the RTX 3080 Around Computex in June 2020

Gaming GeForce RTX 3000 series graphics cards based on the Ampere architecture are said to follow a couple of months later with Computex 2020 being the target. The roll-out will reportedly begin with the introduction of high-end Ampere graphics cards first and will be gradual with Ampere graphics cards replacing their Turing predecessors over several months.

NVIDIA is reportedly planning to announce the RTX 3080 in June with a showcase of the new graphics cards in or just around Computex in early June of next year. Even though the source doesn’t go into details about which specific graphics cards will launch first other than the fact that they will be “high-end”, from a practical point of view it’s more likely that NVIDIA will introduce the RTX 3070 alongside the RTX 3080 and ahead of the RTX 3080 Ti.

This is simply because it’s unlikely that the 7nm process will be mature enough to support the introduction of big Ampere at a consumer level price point so early in Ampere’s life cycle. An Ampere Titan release followed by an RTX 3080 Ti later is more likely. Other GeForce 3000 series graphics cards targeting the mid-range such as the 3060 and 3050 are expected to follow the RTX 3080 and RTX 3070 a couple of months later in the summer.

NVIDIA Ampere RTX 3000 Series – Performance & Pricing ( Rumors )

RTX 3000 series Ampere graphics cards are rumored to bring substantially better ray tracing and rasterization performance. They are also rumored to feature bigger frame buffers cross the lineup as compared to Turing. They will also allegedly have slightly higher clock speeds and lower TDPs compared to their 2000 series Turing counterparts.

In terms of pricing the new cards are rumored to be cheaper in the high-end, specifically the RTX 3080 and 3080 Ti will allegedly be priced more attractively than the RTX 2080 and 2080 Ti.

This RGB memory kit runs faster in certain Gigabyte Aorus motherboards

In certain Aorus motherboards, Gigabyte’s newest DDR4-3600 kit can boost to 3733MHz.

Gigabyte has added an interesting feature to its newest Aorus RGB memory kit, one that allows it to run faster in certain motherboards with minimal fuss. It’s called “Aorus Boost,” and it gives users up to a 4 percent bump in performance.

Even without the boost feature, Gigabyte’s new 16GB (2x8GB) DDR4-3600 memory kit is the fastest one it has released so far. At 3600MHz, the timings are configured at 18-19-19-39, according to the RAM’s main product page and Gigabyte’s press release. (Curiously, the specifications section lists slightly higher timings—19-19-19-43 on Intel platforms and 20-19-19-43 on Ryzen configs, so go figure.)

It’s the boost feature that is somewhat interesting, though. 

“Many users often look to overclock their RAM and tweak their timings for higher memory frequencies and better performance. Memory overclocking can be a hit or miss though with users often spending lots of time trying to unlock better memory performance to little effect as their overclocked memory performance falls short of their expectations,” Gigabyte explains.

This is where Aorus Boost comes into play. On certain Aorus X570 and Z390 motherboards, there is an Aorus Memory Boost setting in the BIOS that will apply the necessary settings to increase the frequency up to 3733MHz. Compatible motherboards include the following models:

Z390 Aorus Xtreme Waterforce 5G

Z390 Aorus Xtreme Waterforce

Z390 Aorus Xtreme

Z390 Aorus Master

Z390 Aorus Ultra

Z390 Aorus Pro WiFi

X570 Aorus Xtreme

X570 Aorus Master

X570 Aorus Ultra

X570 Aorus Pro WiFi

X570 Aorus Pro

X570 I Aorus Pro WiFi

It’s not clear what effect this has on timings; I presume they get loosened a touch. Outside of the boost feature, Gigabyte says this kit is built on a “10-layer, sophisticated PCG that ensures the stability and performance” of the memory chips, which are “100 percent sorted and tested.” Binning is a common practice on the best RAM kits, though, so this is not unique to Gigabyte.

There are two versions of this kit available. Both sport a pair of 8GB modules, but the second version also includes a couple of “demo modules.” These are dummy modules for people who want to fill all their DIMM slots with RGB lighting, without buying more RAM.

Gigabyte did not say when this RAM will be available or how much the two versions will cost. As a point of reference, Gigabyte’s Aorus RGB 16GB DDR4-3200 memory kit with demo modules sell for $169.99 on Amazon.

That said, if you are in the market for RAM (or pretty much anything else), it might be worth waiting a week to see what of Black Friday deals emerge.

Chinese SSD Manufacturer Races Forward to PCIe 5.0

According to a report by Chinese media IThome, Jiangsu Huacun Electronic Technology has demoed its PCIe 5.0 controller at the 2019 Nantong New Generation Information Technology Expo. The memory and storage manufacturer is optimistic that the controller will be in mass-production by the end of 2020.

It was only this year that computer hardware started to arrive with support for the PCIe 4.0 standard. Presently, PCIe 4.0 is still limited to AMD’s current Ryzen desktop CPUs, Ryzen Threadripper HEDT (high-end desktop) CPU line and Epyc server processor portfolio. Intel has yet to board the PCIe 4.0 train, but is expected to do so soon.

Little is known about the HC9001 PCIe 5.0 controller. It’s reportedly produced with the 12nm manufacturing process by the China National Research Institute and Jiangsu Huacun Electronic Technology. The HC9001, which is China’s first domestic PCIe 5.0 controller, is gaining a lot of buzz, so perhaps we’ll get more information as time progresses.

The PCIe 5.0 standard promises to deliver a throughput up to 128 GBps over a x16 slot. This is double the bandwidth of what the PCIe 4.0 interface can do. At this point, PCIe 5.0 would be overkill for the average consumer, considering that today’s mainstream graphics cards and SSDs aren’t fast enough to use PCIe 4.0 to its max potential. However, PCIe 5.0 will certainly be a welcomed progression in the cloud computing, big data, AI and 5G industries.

An unconfirmed Intel roadmap shows the chipmaker adopting the PCIe 5.0 standard with its forthcoming Sapphire Rapids enterprise-grade processors in 2021. So far, we haven’t heard anything about PCIe 5.0 on the AMD front.

ASRock X570 Creator Review: Aqua-Free Thunderbolt 3

OUR VERDICT

What looks on paper to be a half-priced version of the X570 Aqua is really a completely different product. But the X570 Creator’s impressive feature set still makes it worth its price to those in the market for integrated Thunderbolt 3 and 10GbE.

FOR

Integrated Thunderbolt 3 with DisplayPort passthrough

10GbE, Gigabit Ethernet & 802.11ax Wi-Fi

Excellent CPU overclocking at moderate voltage levels

AGAINST

Half the SATA ports are on PCIe 2.0 controllers

All three PCIe x1 slots are 2.0 version

All PCIe 2.0 devices share bandwidth on a single-lane hub

As much as bells and whistles are impressive, ASRock’s recently released X570 Aqua is extremely expensive at $1,000. But the $500 X570 Creator ($480, with a $20 rebate as of this writing) is in many ways a similar board that’s half the price. It doesn’t have the exclusivity or open-loop cooling of the flagship Aqua, but it still has a full-fledged feature set. 

Specifications

Nothing in our features table differentiates this board from the X570 Aqua that went before it, but we still need to dig deeper, as those identical features have been placed on completely different circuit boards.

The lack of a monoblock and color-matched cladding gives the less-costly X570 Creator a completely different look from its feature-matched predecessor. But it does have all the same ports and slots. In fact, the lack of the Aqua’s integrated I/O shield is the only thing that differentiates the Creator’s I/O panel.

The BIOS flash mode button, dual antenna connections for factory-installed 802.11ax module, six USB3 Gen1 and two Thunderbolt 3 ports, Gigabit and 10GbE ports, audio jacks with digital optical output, DisplayPort input and HDMI output are still present. The video input is for Thunderbolt passthrough via external graphics card linking, and the digital audio output supports DTS Connect to encode 5.1-channel streams from live audio sources. The biggest disappointment is that the two Type-A ports that appear color-coded for 10Gbps but are only wired for 5Gbps.

Unlike the X570 Aqua, the Creator fits completely within the confines of standard ATX. We looked around for crowded components and found nothing unusual, though the voltage regulator and socket are both closer to the I/O shield and the chipset is closer to the PCIe slots. The same onboard controllers are in similar positions relative to each other, but with less spacing between them. We even see the same internal DisplayPort pass-through connector, which is used for connecting specific ASRock graphics cards internally to the board’s Thunderbolt 3 outputs, located forward of the I/O audio connections.

Internal power and reset buttons are moved slightly rearward to make room for the X570 Creator’s shortened front edge by moving its legacy PC Speaker/3-pin spaced redundant power LED combo header, compared to the X570 Aqua. Farther up the board we see that the ALC1220 audio codec is moved further down than on the X570 Aqua, and that the ASM1187e PCIe 2.0 x1 to x7 smart switch is now closer to the battery. Builders are welcome to use up to eight SATA devices, but the fact that four of these ports share bandwidth with three x1 slots through that switch’s single-lane 5Gb/s interface might give them pause.

The X570 Creator’s installation kit adds an I/O shield compared to the X570 Aqua, since that one came pre-installed on the motherboard. The same four SATA cables, Wi-Fi antenna, legacy high-bandwidth SLI bridge, and link cable for DisplayPort-to-Thunderbolt passthrough are included.

Software

We played around with ASRock’s A-Tuning long enough to determine that it at least controls our CPU clocks, though its System Info page shows Ryzen Master’s CPU temperature reading as “SB TSI CPU”.

The X570 Creator has no onboard lighting, so its Polychrome Sync app only works for RGB headers and compatible DRAM. Though changes were laggy with our memory, it did eventually work.

Firmware

The tested version of X570 Creator firmware doesn’t have an easy-mode GUI, but its advanced mode is simple to navigate. Its OC Tweaker menu is of particular interest to us, since it includes everything we need to push our low-risk overclocks.

Attempting an all-core overclock means living with the maximum frequency of the worst core, which in our case isn’t very high. As with most boards, attempts to go beyond 4.2 GHz were unsuccessful even at far-higher voltage levels, so we reverted to safe settings.

Ryzen 3000 processors are famed for their memory overclocks but also known for needing sub-optimal memory controller and Infinity Fabric ratios to reach these. Users who don’t want to exceed AMD’s default DDR4-3600 limit for those ratios may instead choose to chase memory performance via optimized timings, for which the X570 Creator includes the full set of controls.

The X570 Creator doesn’t have onboard RGB lighting, and anyone who doesn’t want to install ASRock’s RGB application for Windows can control compatible RGB accessories via the board’s external RGB headers via firmware.

ASRock left the fan mapping graphic out of X570 Creator firmware, but the corresponding numeric values are still there for those who’d rather not use the included software. The default map for the southbridge fan kept it running at 3700 to 5000 RPM, its noise minimized only by its small size.

Overclocking

We’re using the hardware from our first X570 review to compare the X570 Creator to our first four boards, including Fractal Design’s Celsius S24, Gigabyte’s GeForce RTX 2070 Gaming OC 8G and Toshiba’s OCZ RD400 512GB NVMe SSD. Because the previously-reviewed X570 Aqua required a different cooling system, we’ve left it out of our charts.

The X570 Aqua reached DDR4-4000 with all four DIMMs installed, while the X570 Creator only reached DDR4-3800. The Creator still beat the X570-E Gaming and X570 Taichi, but that’s probably due to its newer firmware. It also beats those two boards, and only those two boards, in CPU overclocking, if by a scant 25 MHz.

We overclock to gain performance, but the X570 Creator had the third-worst overclocked memory bandwidth. The slower (data rate) boards were faster (more bandwidth), but we should note that settings over DDR4-3600 can hurt performance due to firmware automatically decoupling the CPU’s Infinity Fabric clock at higher frequencies. This can be overcome via manual overclocking, but typically not at data rates this high. 

3DMark and PCMark

We credit the X570 Creator’s newer firmware for higher CPU scores in 3DMark, but that gain barely made an impact in PCMark Essentials, with a less-than-1% advantage in web browsing. It made a clean sweep of PCMark Digital Content Creation, but came out roughly tied in MS Office.

3D Gaming

The X570 Creator swaps positions at various Ashes settings to take third place overall, but somehow gets a big enough jump in just one of our F1 2017 tests to emerge victorious by a mere 1%.

Mixed Applications

The X570 Creator also showed higher scores on Cinebench, as well as our complete set of timed workloads. If this is what AMD’s new microcode does, our congratulations should go there.

Power, Heat and Efficiency

The X570 Creator throttles our CPU back to 3950 MHz under a fully multithreaded load of Prime95, while the Taichi from the same brand ran up to 4.1 GHz and the Strix X570-E Gaming ran 3950 to 3975 MHz.

Since the CPU power settings of both boards are so similar, it’s important to discuss how the X570 Creator compares to the X570-E Gaming. Both boards have similar CPU thermal readings, though the ROG Strix model has a slightly warmer voltage regulator.

Gigabyte and MSI samples skewed our efficiency results by clocking our CPU down to 3.8 GHz. Even if we adjusted our chart to account for the approximately 4% difference in clock speed, the MEG X570 Ace would still win the efficiency crown.

Final Analysis

One thing to note in our performance-per-dollar chart: The Creator has more features than its competitors. If all you care about is performance and price, that makes the Creator tough to argue for. But you do get more for you money in other ways with this board.

None of the comparison boards have Thunderbolt 3 or 10GbE, though the copper cabling of 2.5GbE give the second network interface of Asus, Gigabyte and MSI broader compatibility.

The cheapest of the comparison boards, ASRock’s own X570 Taichi doesn’t even have a second wired network interface. The X570 Creator’s extra features are probably worth an extra $200 at most, and it costs exactly $200 more.

Competing samples from Asus, Gigabyte, and MSI cost $30 to $60 more, and our previous reviews pointed out how each of these was worth the upgrade. That just makes this a tie from a value perspective, so our best advice is to pick from these the board that best fits your compatibility, efficiency, and overclockability needs.

Intel Claims Xeon Cascade Lake-AP 56 Core CPU Up To 84% Faster Than AMD’s 64 Core EPYC Rome 7742 in Real-World HPC Benchmarks

Intel has posted a series of new benchmarks claiming that their Xeon class Cascade Lake-AP CPUs run much faster than AMD’s 2nd Generation EPYC Rome CPUs. The benchmarks which Intel claims are representative of ‘real-world’ performance in the HPC segment compare Intel’s 2S (dual socket) Xeon Platinum 9282 versus AMD’s EPYCX 7742 (also in a dual-socket configuration).

The Cascade Lake-AP Xeon Platinum 9282 56 Core CPU Claims To Be 84% Faster Than AMD’s EPYC 7742 64 Core CPU in Intel’s ‘Real-World’ HPC – AI Performance Benchmark Suite

The performance metrics for both processors were posted at Medium where Intel also recently published an article about core scaling and the reliance of modern applications on the number of cores available on processors. According to Intel, 8 cores along with sustained frequencies would result in better scaling than say a 12 core or a 16 core chip. Now Intel may have provided a lot of data backing up their findings but the same Intel was reluctant to move beyond 4 cores back in 2017 when AMD was offering their Ryzen chips with up to 8 cores and 16 threads. It’s interesting how suddenly, 8 core processors have become the next-big-thing for Intel’s mainstream lineup and the same would happen with 10th Gen family which is expected to receive even more cores.

For the HPC market, Intel says that – More processor cores add compute, but the overall system or workload performance depends on other factors, including:

· The performance of each core
· Software optimizations leveraging specific instructions
· Memory bandwidth to ensure feeding of the cores
· Cluster-level scaling deployed

Anyway, coming back to the topic, Intel’s latest benchmarks compared the Xeon Platinum 9200 versus the EPYC 7742. The Xeon Platinum is one of the elusive Cascade Lake-AP processors which feature two dies instead of a single monolithic one, stacking up to 56 cores and 112 threads. The chip has a base clock of 2.60 GHz and a boost clock of 3.80 GHz along with 77 MB of cache and a TDP of 400W. The Intel Cascade Lake-AP chips feature 12 memory channels compared to AMD’s 8 memory channels per chip.

The AMD EPYC 7742 is based on a 7nm process node (vs Intel’s 14nm+++) and features 64 cores / 128 threads. The chip has a base clock of 2.25 GHz and a boost clock of 3.4 GHz with 256 MB of L3 cache, 128 PCIe Gen 4 lanes and a TDP of 225W. The pricing plays a huge role too and here we see the EPYC 7742 with a price of $6950 US while the Xeon Platinum 9282 is suggested to have a price between $25K – $50K .

So right off the bat, we can note that this isn’t a fair comparison as not only does Intel’s chip have a higher running TDP but its cost is at least 3.5x higher than the AMD processor. Yes, the EPYC 7742 is AMD’s flagship 2nd Generation Rome processor for servers but even still, this isn’t an apples to apples comparison in any possible way.

Update: ServerTheHome’s Patrick J Kennedy has found out that the GROMACS version being used by Intel is an outdated one that doesn’t utilize the 256-bit wide AVX2 SIMD instruction set that is featured on Zen 2. The GROMACS 2019.3 version was used by Intel in what they are terming as real-world benchmarks, however, the latest version available is 2019.4 which adds support for Zen 2 based EPYC Rome chips like the EPYC 7742 which Intel tested their Xeon Platinum 9282 against. It just goes off to show that even Intel’s ‘Real-World’ benchmarks aren’t indicative of actual product performance and may lead to misleading statements against competitor products. And this won’t be the first time Intel is using misleading benchmarks or statements to downplay the competition. They have termed several important performance metrics used by tech reviewers that are invalid and not indicative of actual product performance while their own performance metrics surely are.

The benchmarks show that the Xeon Platinum 9282 delivers an average performance increase of 31%, going as high as up to 84%. There are several HPC specific applications shown which Intel claims are representative of real-world performance metrics in the server market. Dissecting each application reveal the breakdown of performance in each individual workload for the benchmark and in the case of the Manufacturing application (ANSYS Fluent Workload), Intel has a 13% average performance uplift over AMD’s EPYC Rome chip. Intel also claims that having AVX-512 onboard the new Xeon chips gives them an edge in several applications such as VASP, NAMD, GROMACS, FSI & LAMMPS.

The HPC segment is broad with varying compute requirements by workload. 56 core Xeon Platinum 9282 ranges from 8% to 84% better performance (31% higher geomean) than AMD’s 64 core Rome-based system (7742) on leading real-world HPC workloads across manufacturing, life sciences, financial services and earth sciences(2).

Some of the applications and results are shown above are a geomean of several specific workloads, all with different characteristics and sensitivities. Drilling into the details of these workloads provides further insight into performance. For example, Xeon Platinum 9282 leads AMD Rome 7742 by 13% on a geomean of 14 ANSYS Fluent workloads. Across those 14 different CFD simulations, Xeon’s results range from 2% lower to 36% higher.

Medium (Intel)

Intel further goes on to claim that Xeon Platinum 9200 series processors offer a lower TCO (Total Cost of Ownership). Since the performance of Xeon Platinum 9200 series is higher, you’d have to require a fewer number of nodes which should drive down the node acquisition cost, lower fabric, switching and cabling cost. It is also mentioned that while the Xeon-AP has a higher TDP and power requirement than AMD’s EPYC Rome (225W vs 400W), it should be offset by the lower number of nodes required to reach the same performance.

Aside from raw compute power, memory bandwidth is also highlighted as the main performance measurement factor and surprisingly, major industry players are already evaluating replacing their existing Intel-based systems with EPYC processors. Just a day before Intel published their report on performance, it was revealed that Netflix may soon be switching to AMD’s EPYC based platform as the TCO is similar but the EPYC solution may actually offer higher bandwidth than an Intel Xeon based system.

There will be a lot more action next-year in the server department as AMD will launch their energy efficient 7nm+ EPYC Milan CPUs to tackle both, Intel 14nm Cooper Lake and 10nm Ice Lake lineups, simultaneously.

MSI is dishing up boot-shortening BIOS updates for its X570 motherboards

As the company previously promised, it has begun rolling out firmware updates based on AMD’s latest microcode (AGESA 1.0.0.4), for its X570 motherboards, some of which are on our list of the best gaming motherboards. According to MSI, these BIOS updates bring with them over 100 improvements and fixes, and can shorten boot times as well.

“The latest 1.0.0.4 Patch B BIOS (SMU v46.54) comes with a massive improvement at all points concerning debugs and optimization for the previous version. There is a loooooooooong list regarding improvements and amendments. The most attractive one to me is optimized system boot up procedure that allows us to shorten boot time and less waiting,” MSI stated in a blog post last week.

Compared to BIOS releases based on the previous microcode, MSI says it has seen boot times improve by as much as 8 seconds, after clearing the CMOS.

It’s always a good idea to back up your data before flashing the firmware, in case something goes wrong. Also take note of your BIOS settings—you may need to adjust the settings after flashing to maintain that overclock you spent time fine tuning, to avoid a broken RAID array, and avoid other potential problems.

Whether the allure of faster boots is enough to warrant a BIOS flash, you’ll have to decide that for yourself—SSDs are a boon for boot times and you may find Windows (or Linux) loads fast already.

Beyond faster boots, AMD’s 1.0.0.4 microcode is also supposed to increase all-core boost clocks on Ryzen processors, though to what extent is not known.

Looking ahead, MSI says it will have new BIOS updates available for its 400 Max series motherboards around the middle of November, and for all 400 and 300 series motherboards by the end of next month. Other motherboard makers are expected to release new BIOS updates soon as well.

Aorus’ New PCIe 4 SSD Gobbles Up 16 Lanes of PCIe, Might Make Your GPU Jealous

Even though it’s not really a secret anymore, Gigabyte now formally announced this week the Aorus Gen4 AIC SSD 8TB. This is a PCIe add-in-card which houses four 2TB PCIe 4.0 x4 NVMe SSDs, which turns it into, for lack of better words, an absolute storage monster.

Add those numbers up, and you’ve got an 8TB NVMe SSD that runs over the new PCIe 4.0 standard with access to 16 lanes. Gigabyte claims that it should manage a throughput (read and write) of about 15,000 MBps. The unit is rated at 430,000 random read IOPS and 440,000 random write IOPS.

Putting Things in Perspective

Let’s just put this into perspective. The typical 2.5-inch SATA SSD has a throughput of around 550 MBps, and if you’ve got a fancy PCIe 3.0 x4 NVMe M.2 SSD, you might have a read speed of about 3,500 MBps. This makes Gigabyte’s new AIC almost 30 times as fast as the SATA SSDs many users have.

Gigabyte’s new AIC SSD does need some cooling, though. It comes with a blower-style cooler that pulls air in and exhausts it out the back of the card, outside the PC case. On its way, it cools a copper heatsink placed over the quadruplet of M.2 SSDs installed inside the unit. Inside Gigabyte also fitted the unit with eight temperature sensors, which can be monitored from within the Aorus Storage Manager software. Interestingly, this SSD is one of the few Aorus-branded products that doesn’t come with RGB lighting (though we doubt that’s a primary concern of anyone looking at this offering). 

The M.2 SSDs inside the unit are built with Toshiba BiSC4 96-layer 3D TLC NAND, which enable a transfer rate of up to 800 MTps. The NAND on each SSD is brought together by a Phison PS5016-E16 controller. Durability is rated at 3,600 TBW per 2TB SSD. Because we expect this to be a rather costly purchase that any user would want to last, it’s fortunate Gigabyte offers a 5-year warranty with it.

This all sounds good, but is there is a caveat. As noted, the Aorus Gen4 AIC SSD 8TB runs on PCIe 4.0. So the only platform you can currently use this with is an an AMD X570 motherboard paired with an AMD Ryzen 3000 CPU. The Ryzen 3000 CPUs have 24 PCIe lanes – four of which are downstream to the chipset. Consequently, from the 20 lanes left, you’d eat up 16 with the Aorus SSD. That leaves only 4 lanes, which may be enough for a graphics card (given that PCIe 4.0 is effectively twice as fast as PCIe 3.0), but the PCIe 4.0 interface is so new it’s hard to know for sure without testing. 

Thankfully, the rumor mill has been working hard and the TRX40 platform for AMD’s Ryzen 3000 Threadrippper CPUs is set to launch next month. Given that the CPUs will carry similar specs as AMD’s Epyc line of CPUs, which have 128 PCIe 4.0 lanes, we’re confident that the Ryzen 3000 Threadripper will have a sufficient number of lanes for utilizing Gigabyte’s new Aorus Gen4 AIC SSD 8TB.

No pricing has been revealed yet.

AMD’s comments on CPU design reflect it is in a very different place than Intel right now

While Intel is making investments to recapture process leadership, AMD says it’s more focused on architecture.

AMD just wrapped up a quarter in which it generated the highest revenue since 2005, and that in and of itself is impressive. The real story is not in the sales numbers, though. What I find more interesting than anything else, as it pertains to AMD’s current situation, is how it finds itself operating in a much different mindset than rival Intel, which for years offered up the most advanced processor technology on the market.

That is no longer the case. Yes, the newly released Core i9-9900KS is the best CPU for gaming, and before that, the regular Core i9-9900K held the distinction. But underneath the hood, it is yet another iteration of Intel’s 14-nanometer lithography (14nm++, if you’re keeping count), which dates all the way back to Intel’s 5th generation Broadwell parts released in 2014. And if you look beyond gaming, AMD is swinging more cores in the mainstream market, to hammer multi-threaded workloads. The gap will widen further when AMD’s 16-core/32-thread Ryzen 9 3950X arrives next month.

AMD’s resurgence began with the introduction of its Zen architecture. However, now in its third generation, Zen 2 (which follows Zen and Zen+) is the first to employ a 7nm process node.

Comparing process nodes from one company to another is not exactly an apples-to-apples affair, but 7nm is certainly cutting edge, and it has arrived in earnest. Meanwhile, Intel only recently (and finally) began volume shipments of its 10nm Ice Lake processors, though so far they are only available in laptops.

Process node is not everything, but it’s important. It’s also where AMD and Intel find themselves at very different places, not just in what’s currently shipping, but in where each company’s focus is, at the moment.

I bring this up because of comments made by AMD CEO Dr. Lisa Su during an earnings call with investors. She was asked to break down how much of AMD’s current advantage is really related to process technology versus the underlying architecture.

“The way I would answer that question is, we’ve made a set of choices, and the set of choices include process technology, they include architecture, our chiplet architecture, they include sort of our overall system architecture. And I think we’ve made it set of good choices,” Dr. Su said (a full transcript is available at SeekingAlpha).

“Going forward, we are not relying on process technology as the main driver. We think process technology is necessary. It’s necessary to be sort of at the leading edge of process technology. And so, today, 7-nanometer is a great node, and we’re getting a lot of benefit from it. We will transition to the 5-nanometer node at the appropriate time and get great benefit from that as well. But we’re doing a lot in architecture. And I would say, that the architecture is where we believe the highest leverage is for our product portfolio going forward.”

In a sense, Dr. Su is almost downplaying its process technology lead. It’s a luxury that would have been difficult to fathom before Zen arrived, and in particular Zen 2. But here we are, with AMD shipping 7nm CPUs and talking about architecture being where its focus is at going forward, rather than relying on process technology to be its main driver.

Contrast that sentiment with Intel, which recently said it’s “investing to recapture process leadership going forward.” I’m sure Intel would say architecture is important too, but unlike AMD, getting back on a consistent cadence with regards to its process node is where Intel is focused on.

“We’re ramping a multitude of products. We have increased confidence in 5-nanometer. And as we mentioned for 7 and 5 getting back to a two-and-a-half, two year cadence is what we’re focused on and we’re confident in the future,” Intel CEO Bob Swan said last week.

I fully expect Intel to rebound on the technology side (it already leads in sales, revenue, and market share). Getting to 10nm has been tough, largely because Intel’s goals were too aggressive at the outset. But I think things will go smoother in the next few years. We’ll see.

Regardless of how it plays out, AMD is in a good spot. Perhaps just as importantly, AMD is exuding the kind of genuine confidence that wasn’t really there in the past. AMD had that swagger in the early Athlon 64 / X2 days, and now it’s back, only this time AMD seems in a better position, having divested itself of the manufacturing side of the business. In retrospect, that move is proving to have been extremely beneficial.