Author:
Manufacturer: AMD

Memory Matters

Memory speed is not a factor that the average gamer thinks about when building their PC. For the most part, memory performance hasn't had much of an effect on modern processors running high-speed memory such as DDR3 and DDR4.

With the launch of AMD's Ryzen processors, last year emerged a platform that was more sensitive to memory speeds. By running Ryzen processors with higher frequency and lower latency memory, users should see significant performance improvements, especially in 1080p gaming scenarios.

However, the Ryzen processors are not the only ones to exhibit this behavior.

Gaming on integrated GPUs is a perfect example of a memory starved situation. Take for instance the new AMD Ryzen 5 2400G and it's Vega-based GPU cores. In a full Vega 56 or 64 situation, these Vega cores utilize blazingly fast HBM 2.0 memory. However, due to constraints such as die space and cost, this processor does not integrate HBM.

DSC04643.JPG

Instead, both the CPU portion and the graphics portion of the APU must both depend on the same pool of DDR4 system memory. DDR4 is significantly slower than memory traditionally found on graphics cards such as GDDR5 or HBM. As a result, APU performance is usually memory limited to some extent.

In the past, we've done memory speed testing with AMD's older APUs, however with the launch of the new Ryzen and Vega based R3 2200G and R5 2400G, we decided to take another look at this topic.

For our testing, we are running the Ryzen 5 2400G at three different memory speeds, 2400 MHz, 2933 MHz, and 3200 MHz. While the maximum supported JEDEC memory standard for the R5 2400G is 2933, the memory provided by AMD for our processor review will support overclocking to 3200MHz just fine.

Continue reading our look at memory speed scaling with the Ryzen 5 2400G!

G.Skill Launches 4700 MHz Trident Z RGB DDR4 Memory

Subject: General Tech, Memory | February 11, 2018 - 04:45 PM |
Tagged: G.Skill, Trident Z RGB, ddr4, Samsung, samsung b-die, xmp

G.Skill will soon be upgrading its Trident Z RGB line of DDR4 DIMMs with a 16 GB kit capable of running at 4700 MHz. With the claimed fastest commercial kit of RGB-equipped memory modules, the new 2 x 8 GB kit uses Samsung B-die ICs and supports XMP 2.0 memory profiles. The super-fast memory kit has been in development for quite a while and is slated for availability in Q2 2018.

GSkill Trident Z RGB.jpg

G.Skill has managed to tighten the timings on its 4700 MHz kit to CL19-19-19-39 while needing only 1.45V which is nice to see. The company has reportedly validated the new memory using a MSI Z370 Gaming Pro Carbon AC motherboard and Intel i7 8700k processor. G.Skill notes that the new kit is notable because it is the first retail kit to hit 4700 MHz as well as the first memory kit with RGB LEDs to hit that lofty memory speed. Corsair comes close at 4600 MHz with its 16 GB Vengeance LPX DDR4 kit at 15-15-15-36 which will set you back a cool $589.99 MSRP.

I am curious on the overclocking headroom on these modules actually (heh). G.Skill is reportedly using highly screened B-dies so maybe the 5,000 MHz its other kits have hit (when overclocked) would be possible. I would like to see AMD’s Infinity Fabric performance at that point when it is not being held back by memory speed especially where its upcoming APUs are concerned. On the Intel side of things, I think tighter timings are preferrable (after a certain threshold of acceptable speed of course) when pursuing the best performance so a "slower" 3600 to 4600 MHz kit at CL15 or lower might be a better buy. In any case, memory continues to be pricey, and I would uess G.Skill's new kit will hit at least $600 MSRP.

G.Skill is not yet talking pricing on these modules, but they aren’t going to be cheap. We should know more in a couple of months as we enter the second quarter.

Also read:

Source: G.Skill

SK Hynix Launches Its 8Gb GDDR6 Memory Running at 14 Gbps

Subject: Graphics Cards, Memory | January 24, 2018 - 11:04 PM |
Tagged: SK Hynix, graphics memory, gddr6, 8gb, 14Gbps

SK Hynix recently updated its product catalog and announced the availability of its eight gigabit (8 Gb) GDDR6 graphics memory. The new chips come in two SKUs and three speed grades with the H56C8H24MJR-S2C parts operating at 14 Gbps and 12 Gbps and the H56C8H24MJR-S0C operating at 12 Gbps (but at higher voltage than the -S2C SKU) and 10 Gbps. Voltages range from 1.25V for 10 Gbps and either 1.25V or 1.35V for 12 Gbps to 1.35V for 14 Gbps. Each 8 Gb GDDR6 memory chip holds 1 GB of memory and can provide up to 56 GB/s of per-chip bandwidth.

SK Hynix logo.jpg

While SK Hynix has a long way to go before competing with Samsung’s 18 Gbps GDDR6, its new chips are significantly faster than even its latest GDDR5 chips with the company working on bringing 9 Gbps and 10 Gbps GDDR5 to market. As a point of comparison, its fastest 10 Gbps GDDR5 would have a per chip bandwidth of 40 GB/s versus its 14 Gbps GDDR6 at 56 GB/s. A theoretical 8GB graphics card with eight 8 Gb chips running at 10 Gbps on a 256-bit memory bus would have maximum bandwidth of 320 GB/s. Replacing the GDDR5 with 14 Gbps GDDR6 in the same eight chip 256-bit bus configuration, the graphics card would hit 448 GB/s of bandwidth. In the Samsung story I noted that the Titan XP runs 12 8 Gb GDDR5X memory chips at 11.4 Gbps on a 384-bit bus for bandwidth of 547 GB/s. Replacing the G5X with GDDR6 would ramp up the bandwidth to 672 GB/s if running the chips at 14 Gbps.

Theoretical Memory Bandwidth
Chip Pin Speed Per Chip Bandwidth 256-bit bus 384-bit bus 1024-bit (one package) 4096-bit (4 packages)
10 Gbps 40 GB/s 320 GB/s 480 GB/s    

12 Gbps

48 GB/s 384 GB/s 576 GB/s    
14 Gbps 56 GB/s 448 GB/s 672 GB/s    
16 Gbps 64 GB/s 512 GB/s 768 GB/s    
18 Gbps 72 GB/s 576 GB/s 864 GB/s    
HBM2 2 Gbps 256 GB/s     256 GB/s 1 TB/s

GDDR6 is still a far cry from High Bandwidth Memory levels of performance, but it is much cheaper and easier to produce. With SK Hynix ramping up production and Samsung besting the fastest 16 Gbps G5X, it is likely that the G5X stop-gap will be wholly replaced with GDDR6 and things like the upgraded 10 Gbps GDDR5 from SK Hynix will pick up the low end. As more competition enters the GDDR6 space, prices should continue to come down and adoption should ramp up for the new standard with the next generation GPUs, game consoles, network devices, ect. using GDDR6 for all but the highest tier prosumer and enterprise HPC markets.

Also read:

Samsung Begins Mass Production Of 18 Gbps 16-Gigabit GDDR6 Memory

Subject: Memory | January 18, 2018 - 12:34 AM |
Tagged: Samsung, graphics memory, graphics cards, gddr6, 19nm

Samsung is now mass producing new higher density GDDR6 memory built on its 10nm-class process technology that it claims offers twice the speed and density of its previous 20nm GDDR5. Samsung's new GDDR6 memory uses 16 Gb dies (2 GB) featuring pin speeds of 18 Gbps (gigabits-per-second) and is able to hit data transfer speeds of up to 72 GB/s per chip.

Samsung GDDR6_PhotoFs.png

According to Samsnug, its new GDDR6 uses a new circuit design which allows it to run on a mere 1.35 volts. Also good news for Samsung and for memory supply (and thus pricing and availability of products) is that the company is seeing a 30% gain in manufacturing productivity cranking out its 16Gb GDDR6 versus its 20nm GDDR5. 

Running at 18 Gbps, the new GDDR6 offers up quite a bit of bandwidth and will allow for graphics cards with much higher amounts of VRAM. Per package, Samsung's 16Gb GDDR6 offers 72 GB/s which is twice the density, pin speed, and bandwidth than that of its 8Gb GDDR5 running at 8Gbps and 1.5V with data transfers of 32 GB/s. (Note that SK Hynix has announced it plans to produce 9Gbps and 10Gbps dies which max out at 40 GB/s.) GDDR5X gets closer to this mark, and in theory is able to hit up to 16 Gbps per pin and 64 GB/s per die, but so far the G5X used in real world products has been much slower (the Titan XP runs at 11.4 Gbps for example). The Titan XP runs 12 8Gb (1GB) dies at 11.4 Gbps on a 384-bit memory bus for maximum memory bandwidth of 547 GB/s. Moving to GDDR6 would enable that same graphics card to have 24 GB of memory (with the same number of dies) with up to 864 GB/s of bandwidth which is approaching High Bandwidth Memory levels of performance (though it still falls short of newer HBM2 and in practice the graphics card would likely be more conservative on the memory speeds). Still, it's an impressive jump in memory performance that widens the gap between GDDR6 and GDDR5X. I am curious how the GPU memory market will shake out in 2018 and 2019 with GDDR5, GDDR5X, GDDR6, HBM, HBM2, and HBM3 all being readily available for use in graphics cards and where each memory type will land especially on the mid-range and high-end consumer cards (HBM2/3 still holds the performance crown and is ideal for the HPC market).

Samsung is aiming its new 18Gbps 16Gb memory at high performance graphics cards, game consoles, vehicles, and networking devices. Stay tuned for more information on GDDR6 as it develops!

Also read:

Source: Samsung

Samsung Mass Producing Second Generation "Aquabolt" HBM2: Better, Faster, and Stronger

Subject: Memory | January 12, 2018 - 05:46 PM |
Tagged: supercomputing, Samsung, HPC, HBM2, graphics cards, aquabolt

Samsung recently announced that it has begun mass production of its second generation HBM2 memory which it is calling “Aquabolt”. Samsung has refined the design of its 8GB HBM2 packages allowing them to achieve an impressive 2.4 Gbps per pin data transfer rates without needing more power than its first generation 1.2V HBM2.

SAMSUNG-HBM2_C.jpg

Reportedly Samsung is using new TSV (through-silicon-via) design techniques and adding additional thermal bumps between dies to improve clocks and thermal control. Each 8GB HBM2 “Aquabolt” package is comprised of eight 8Gb dies each of which is vertically interconnected using 5,000 TSVs which is a huge number especially considering how small and tightly packed these dies are. Further, Samsung has added a new protective layer at the bottom of the stack to reinforce the package’s physical strength. While the press release did not go into detail, it does mention that Samsung had to overcome challenges relating to “collateral clock skewing” as a result of the sheer number of TSVs.

On the performance front, Samsung claims that Aquabolt offers up a 50% increase in per package performance versus its first generation “Flarebolt” memory which ran at 1.6Gbps per pin and 1.2V. Interestingly, Aquabolt is also faster than Samsung’s 2.0Gbps per pin HBM2 product (which needed 1.35V) without needing additional power. Samsung also compares Aquabolt to GDDR5 stating that it offers 9.6-times the bandwidth with a single package of HBM2 at 307 GB/s and a GDDR5 chip at 32 GB/s. Thanks to the 2.4 Gbps per pin speed, Aquabolt offers 307 GB/s of bandwidth per package and with four packages products such as graphics cards can take advantage of 1.2 TB/s of bandwidth.

This second generation HBM2 memory is a decent step up in performance (with HBM hitting 128GB/s and first generation HBM2 hitting 256 GB/s per package and 512 GB/s and 1 TB/s with four packages respectively), but the interesting bit is that it is faster without needing more power. The increased bandwidth and data transfer speeds will be a boon to the HPC and supercomputing market and useful for working with massive databases, simulations, neural networks and AI training, and other “big data” tasks.

Aquabolt looks particularly promising for the mobile market though with future products succeeding the current mobile Vega GPU in Kaby Lake-G processors, Ryzen Mobile APUs, and eventually discrete Vega mobile graphics cards getting a nice performance boost (it’s likely too late for AMD to go with this new HBM2 on these specific products, but future refreshes or generations may be able to take advantage of it). I’m sure it will also see usage in the SoCs uses in Intel’s and NVIDIA’s driverless car projects as well.

Source: Samsung

Studying the effect of frequency and timing for Coffee

Subject: Memory | October 18, 2017 - 04:20 PM |
Tagged: coffee lake, i7 8700k, Intel

The performance of AMD's Ryzen chips depend heavily on the frequency of the RAM installed thanks to how Infinity Fabric works.  TechPowerUp decided to see how sensitive Intel's Coffee Lake processors are, testing the performance with RAM speeds from 2133MHz up to 4000MHz as well as modifying the timings.  Not to spoil the results for you, we can reveal something else their tests revealed, G.SKILL's Trident Z DDR4-3866 16GB kit is impressively flexible, they were stable at 15 different combinations of timings and frequencies.  Check out the full results to discover the sweet spot.

title.jpg

"We take a close look at memory speeds, latencies and command rate on Intel's latest Core i7-8700K with Z370. Scenarios tested include fail-safe 2133 MHz, the platform default of 2666 MHz and overclocked memory speeds from 3000 MHz to 4000 MHz - at various timings."

Here are some more Memory articles from around the web:

Memory

 

Source: TechPowerUp

Corsair Overclocks With a Vengeance, Launches DDR4 4600 MHz Memory Kit

Subject: Memory | September 21, 2017 - 11:46 AM |
Tagged: X399, ryzen, overclocking, Intel X299, ddr4-4600, ddr4, corsair

Corsair has launched a new Vengeance LPX DDR4 memory kit that is capable of hitting 4600 MHz at 1.5 volts. The new kit is a 16GB (2 x 8GB) kit that Corsair reportedly co-developed with AsRock for fine tune using their X299 OC Formula motherboard. The DDR4 kit is made using hand sorted Samsung B-dies and it supports Intel XMP 2.0 standards allowing it to clock at 4600 MHz with a single setting change in the UEFI.

Corsair Vengeance LPX 4600 MHz.jpg

The Vengeance LPX DIMMs run with CAS timings of 19-26-26-46 and need only 1.5V to clock at 4600 MHz. This kit will be ideal for Intel’s X299 as well as AMD’s X399 platforms. While Ryzen and Threadripper platforms may need a bit more tweaking to get working, they would benefit the most from the higher clocked memory allowing the Infinity Fabric to clock higher.

Being one of the highest factory clocked DIMMs, they come at a cost. The new RAM kit (CMK16GX4M2F4600C19) is available now for $549.99 with a lifetime limited warranty.

For something a bit more tame, earlier this week Corsair launched a 2 x 8GB kit (CMK16GX4M2F4500C19) clocked at 4500 MHz with CL19-19-19-39 timings (at 1.45V) that is also available now for $479.99 MSRP. Enthusiasts might be better off buying the cheaper kit and overclocking them (though not guaranteed and might need a bit more than 1.5V) while workstation and enterprise customers with corporate expense accounts can opt for the more expensive but factory clocked 4600 MHz kit.

At time of writing the new kits were not up on Amazon yet, but they should be shortly. You can find the cheaper 4500 MHz kit on Corsair's web store but it is listed at $504.99 currently. If you wait a bit, that price should go down closer to MSRP as other retailers put up their listings.

Source: TechPowerUp

G.SKILL Announces New DDR4 for AMD Ryzen Threadripper

Subject: Memory | August 10, 2017 - 03:31 PM |
Tagged: Zen, Threadripper, ryzen, amd, G.Skill, flare x, quad channel

G.SKILL have launched several new kits specifically designed for Threadripper systems, all under the name of Flare X.   There are three 32GB kits and a single massive 128GB kit to choose from, all quad channel and all tested for compatibility with Threadripper.

image004.jpg

unnamed.png

Taipei, Taiwan (10 Aug 2017) – G.SKILL International Enterprise Co., Ltd., the world’s leading manufacturer of extreme performance memory and gaming peripherals, announces all-new DDR4 specifications and expanding the Flare X series, designed for AMD processors and platforms. Compatible with the new Ryzen™ Threadripper™ processors and AMD X399 chipset motherboards, these new DDR4 specifications are designed to achieve high frequency at DDR4-3600MHz 32GB (8GBx4), as well as a massive total capacity at DDR4-2933MHz 128GB (16GBx8). Included in the mix of new quad-channel DDR4 memory kits are DDR4-3200MHz CL14 32GB (8GBx4) and DDR4-3466MHz CL16 32GB (8GBx4).

Ultra-High Frequency Flare X Series Memory Kits at DDR4-3600MHz 32GB (8GBx4)
With improved overclocking performance on the latest AMD Ryzen™ Threadripper™ processors on the X399 chipset, G.SKILL is announcing the DDR4-3600MHz CL16-18-18-38 with 32GB (8GBx4) total capacity running in quad-channel mode, under the Flare X series. Tested for maximum stability, this kit’s frequency speed marks the fastest memory kit ever released thus far for an AMD platform.

Massive Kit Capacity, No Compromises: DDR4-2933MHz 128GB (16GBx8)
One of the advantages introduced by the AMD X399 platform is the increase to 8 memory slots on AMD platforms, allowing the support for massive 128GB capacity kits running in quad-channel mode. Tested using the highest standards for memory stability on AMD Ryzen™ Threadripper™ platforms, G.SKILL announces the Flare X series DDR4-2933MHz CL14-14-14-34 128GB (16GBx8) memory kit running at 1.35V, perfect for systems requiring high-capacity, high-bandwidth memory kits.

Source: G.SKILL

AMD Will Sell Wraith Max CPU Cooler Separately

Subject: Memory | August 6, 2017 - 11:41 AM |
Tagged: wraith max, Wraith, ryzen, fm2, amd, AM4

Amidst all the big AMD announcements recently, the company quietly revealed that it would begin selling the Wraith Max CPU cooler separately at retail. The Wraith Max heatsink and fan was previously only available in OEM systems and in boxed SKUs of the highest end Ryzen processors (mainly the 1700X and 1800X). The cooler is a refreshed and upgraded version of the company’s original Wraith cooler that measures 105 x 105 x 85mm and features a boxy horizontal cooler with a copper baseplate and heatpipes with a shrouded 92mm fan along with a RGB LED ring around the fan that can be controlled via motherboard software.

AMD Wraith Max CPU Cooler.png

The Wraith Max is rated at 140W TDP and is connected to the system using a fan header and USB (for controlling the lighting). AMD further rates the cooler at a fairly quiet 38 dBA. The Wraith Max supports all of the usual AMD sockets including AM4, AM3, and FM2 (no Threadripper support of course heh), but there is no official support for Intel sockets.

The Wraith Max cooler will retail for $59 USD. I have been keeping an eye on the usual online retailers and have not yet seen it listed, but it should be available soon. Hopefully there will be more reviews of the cooler now that it is a retail product on its own, and maybe we can get Sebastian to take a look at it and compare it to the original Wraith cooler (and his usual lineup of course) he reviewed last year.

Source: AMD

G.Skill Readies DDR4-4400 Kits for Intel's X299 HEDT Platform

Subject: Memory | June 29, 2017 - 02:03 PM |
Tagged: x299, trident z, samsung 8Gb, overclocking, G.Skill, ddr4

G.Skill recently announced new DDR4 memory kits for the Intel X299 HEDT platform. The new kits include a dual channel DDR4 4400 MHz kit for Kaby Lake X and a quad channel DDR4 4200 MHz kit for Skylake X. The dual channel kit is available under the company’s Trident Z RGB and Trident Z Black brands depending on whether you want RGB lighting or simple black heatspreaders. The quad channel DDR4-4200 kit is only available in non-RGB Trident Z modules.

Trident Z.png

According to G.Skill, all of the new memory kits use Samsung 8Gb dies and feature CAS latencies of 19-19-19-39. The quad channel 4200 MHz DDR4 modules need 1.40V to hit those specifications, and while it is not yet known what the higher clocked dual channel DDR4 4400 MHz kits need to hit CL19 timings I would presume they need a bit more.

GSkill Memory Kits.jpg

The new kits will be available in an 8GB x 2 (16GB) 4400 MHz kit and up to 64 GB (8GB x 8) 4200 MHz kits. Pricing has not yet been announced, but the new RAM kits should be available soon. While Intel processors do not get as much of a boost from increased memory speeds as AMD’s APUs and Ryzen CPUs do, there are still noticeable gains to be had with faster memory though gamers should still prioritize graphics cards and processors over memory when picking parts for a budget build.

Note that since these kits are using Samsung 8Gb ICs, they also have a good chance of working with Ryzen, but check with your motherboard manufacturer and reviews before ponying up the cash. 

 

Source: G.Skill