Month: December 2019

The new XPG Hunter DDR4 modules come in both conventional DIMM and SO-DIMM form factors.

Adata is expanding its memory line with a new XPG Hunter DDR4 series, available for both desktop PCs (conventional DIMMs) and laptops (SO-DIMMs). The latter form factor also applies to some mini-PCs. The best DDR4 memory kits combine higher speeds with tighter timings, and maybe some RGB bling as a bonus.

“The XPG Hunter modules are made with high-quality chips selected through a strict filtering process. They are equipped with the finest PCBs and pass rigid reliability and compatibility tests to ensure longevity and rugged durability, which are vital for overclocking, gaming, and extreme benchmarking,” Adata says.

Otherwise known as binning, the “strict filtering process” Adata references is a common practice on high performance RAM. Companies test individual chips for their ability to hit and maintain certain speeds, and this is one of the things that can separate an enthusiast kit of RAM from a run-of-the-mill kit.

The modules come wrapped in a stylish heatsink, though if you were hoping for RGB lighting, you won’t find it here. As for the specs that matter (and yes, RAM speed and capacity matters for gaming, to an extent), there are 8GB, 16GB, and 32GB options. Here’s the desktop lineup:

8GB DDR4-2666—CL16-18-18, 1.2V

8GB DDR4-3000—CL16-20-20, 1.35V

16GB DDR4-2666—CL16-18-18, 1.2V

16GB DDR4-3000—CL-16-20-20, 1.35V

32GB DDR4-3000—CL-16-20-20, 1.35V

The SO-DIMM variants for laptops and mini PCs come in the same speeds and capacities, but the timings and voltage are a little different. Here’s a look:

8GB DDR4-2666—CL18-18-18, 1.2V

8GB DDR4-3000—CL17-19-19, 1.2V

16GB DDR4-2666—CL18-18-18, 1.2V

16GB DDR4-3000—CL17-19-19, 1.2V

32GB DDR4-3000—CL17-19-19, 1.2V

None of the kits are showing up in retail channels yet. However, Adata provided some MSRP info, saying the 8GB module costs $39.99 in DIMM form and $49.99 in SO-DIMM form, while the 16GB module costs $79.99 in DIMM form and $99.99 in SO-DIMM form.

I assume those numbers apply to the DDR4-2666 kits, but we’ll have to wait and see how things shake out once these modules and kits show up at places like Amazon and Newegg.

According to chatter on the grapevine, Intel’s first-ever DG1 GPU will be just 23% more powerful than Tiger Lake graphics and is currently struggling to meet its TDP envelope of 25W. Oh and notice the rumor tag and its implications, or forever hold thy silence. Intel’s Xe GPUs have captured the imaginations of tech enthusiasts everywhere and practically all of us cant wait for the duopoly to become a triopoly. But if this rumor is true, (and this is the second time we are hearing something along these lines) there might be more than a few hiccups before that dream can become a reality.

Intel’s DG1 GPU struggling to fit inside TDP envelope of 25W, performance numbers alleged for the first time at 23% faster than TGL

The interesting thing about this rumor is that DG1 was obviously only going to be a training vehicle for Intel to cut its teeth into the world of dGPUs but we have previously heard rumors that Intel wasn’t able to get any AIB on board to trial run DG1. While this is understandable considering no AIB would want to be the guinea pig for Intel to learn the ropes, another potential reason for this refusal has emerged: Intel unable to control the TDP envelope of the platform.

Before we go any further, here is what the Reddit thread states:

New news from Intel insiders this weekend. DG1 is not shaping up very well at all.

In terms of performance DG1’s envelope is stated to be Tiger Lake + ~23%.

Thermals/Power is looking to be dismal as Koduri is having a hard time reigning in DG1’s power envelope. Anyone that has followed his GPU designs know that this has been his downfall forever.

Intel does actually have a couple of notebook design wins currently but it was explained to me that these were deals to lend a bit of legitimacy to Koduri’s program, rather than to actual translate to sales/profits. It was stated that Intel would “give these things away,” if it had to.

The real question now is should we expect to see Koduri still be working for Intel after Q3’20? If not, you can expect to never see DG2 make its way to a retail product.

The thoughts of Intel actually being a player with the likes of AMD and NVIDIA in the world of GPUs is slowly slipping away.

While the rumor has other points of note as well, including the fact that the team is not able to get the TDP envelope under control, I personally believe that this is not something that would be cause for concern for Intel at this stage. This early in the game, the aim should be to get a working dGPU – tweaking can wait. The gaming market has already indicated on multiple occasions that TDP is the least of their concerns and frankly – a 25W GPU isn’t something anyone would seriously buy anyways (unless priced at throwaway MSRP).

A discrete GPU with 23% more performance than Tiger Lake would make sense if the pricing is just right and considering Intel has the financial horsepower (ahem) to employ predatory/penetrative pricing in its first run to break into the market, this shouldn’t really surprise anyone. Yet, here we are, the second in a string of rumors claiming Intel has run into serious trouble with its dGPU efforts.

I still maintain this though: I haven’t heard anything so far that should be cause for concern. Building a GPU isn’t a walk in the park. If it was, we wouldn’t have a duopoly right now. Multiple stumbles, hurdles, and heartbreak are expected before you can have a viable product in hand and if anyone has the will and resources necessary to pull this off, its Intel.

PhysX has been Nvidia’s party trick for well over a decade, but it seems that after all these years AMD wants to play catch up on that front too. The red team’s latest announcement is that its new FEMFX library is finally available, which offers deformable physics effects through the GPUOpen set of AMD-built libraries for developers.

The name FEMFX stems from FEM, which stands for Finite Element Method. AMD describes this as “Solid objects are represented as a mesh of tetrahedral elements, and each element has material parameters that control stiffness, how volume changes with deformation, and stress limits where fracture or plastic (permanent) deformation occur.”

The intention is not for all objects to have these properties, but rather in addition to existing solid objects in order to balance out the workload that stems from a scene — having all elements as finite elements would create a very demanding workload. Of course, if developers are so inclined, they could use the FEMFX library to create entire worlds which are completely destructible.

As mentioned, FEMFX is built as a highly-threaded library, which is a great way for AMD to promote its multi-core Ryzen CPUs. The core counts of those chips have been reaching levels where it no longer makes sense for most gamers to dish out for the higher-end models, so adding multi-threaded CPU-based physics to games can be a great way to utilize otherwise idle cores in the higher end Ryzen chips during gaming.

Some of the materials that can be simulated are elastic deformation, denting metal, wood destruction, and even melting. AMD’s full list of features is below.

Elastic and plastic deformation

Implicit integration for stability with stiff materials

Kinematic control of mesh vertices

Fracture between tetrahedral faces

Non-fracturing faces to control shape of cracks and pieces

Continuous collision detection (CCD) for fast-moving objects

Constraints for contact resolution and to link objects together

Constraints to limit deformation

Dynamic control of tetrahedron material parameters

Support for deforming a render mesh using the tetrahedral mesh

Continuing our dive into newcomer Realtek’s SSD controller lineup thanks to several ADATA products, we turn to the ADATA XPG SX6000 Pro, an entry-level NVMe SSD.

As mentioned in our previous SSD review, Realtek is a well-known PC components supplier that is relatively new to the SSD controller business. ADATA has provided Realtek with their most significant SSD design wins, starting with the original SX6000 that used Realtek’s very first SSD controller, the RTS5760. The SX6000 Pro targets the same entry-level NVMe market segment, but switches to the second-generation RTS5763DL controller. That newer controller brings a desperately-needed process shrink to enable an upgraded PCIe link and much better performance and power efficiency.

As a late entrant to the SSD market, Realtek is focusing primarily on NVMe rather than SATA. Given that they also prefer to be the low-cost, high-volume alternative rather than pursue high-end niches, the RTS5763DL is probably the most important SSD controller in their current lineup. This controller has a four-channel NAND interface, four PCIe lanes and no DRAM controller, so at a high level it is on par with the Silicon Motion SM2263XT and the upcoming Phison E13T.

ADATA has paired the Realtek controller with Micron’s 64-layer 512Gbit TLC NAND, organized as four packages with four dies each for our 1TB sample, all on one side of the M.2 card thanks to the small size of the controller and the lack of an extra DRAM package. The use of 512Gb instead of 256Gb dies cuts costs slightly, but is usually avoided on high-end drives because having fewer dies limits the degree of parallelism available to the SSD controller. This isn’t a problem for the 1TB SX6000 Pro since the four-channel controller is a bigger limitation. Since the 256GB and 512GB models of the SX6000 Pro don’t have severely curtailed performance specs, ADATA may have made the smart choice to use the 256Gb dies on at least the smallest model.

For SATA SSDs, the lack of a DRAM cache pretty much always condemns a drive to have extreme low-end performance. For NVMe SSD, going DRAMless doesn’t carry such a harsh penalty and entry-level NVMe SSDs usually have no trouble offering better real-world performance than mainstream SATA SSDs with DRAM caches. This is often achieved with the help of the NVMe Host Memory Buffer feature that lets the SSD borrow a chunk of the host system’s DRAM (64 MB in the case of this drive) to use as a cache for the drive’s internal data structures used for managing the flash memory. However, toggling HMB on and off didn’t produce a meaningful performance difference for any of our tests on the SX6000 Pro, which leaves us questioning whether ADATA and Realtek enabled the feature just to check a box on the spec sheet. We can’t actually directly observe the drive’s use of HMB without a PCIe bus analyzer, but regardless of how (or if) the SX6000 Pro is making use of HMB, the lack of a measurable effect means we’re only bothering to graph the results from our test runs with HMB on, which is the default for any recent OS.

As usual, ADATA bundles the SX6000 Pro with a thin heatspreader that the user can apply if desired, which we never do for regular reviews. ADATA is also selling the same drive as the XPG GAMMIX S5, which has a slightly fancier heatsink pre-installed.

The Competition

We have tested several other entry-level NVMe SSDs that are in direct competition with the SX6000 Pro. The Mushkin Helix-L uses the Silicon Motion SM2263XT controller, another DRAMless 4-channel design with HMB support. The Intel 660p (and the newer 665p) use the SM2263 with DRAM, and use QLC NAND instead of TLC. The Toshiba (now Kioxia) BG4 doesn’t currently have a retail version, but it’s a popular OEM SSD that likewise is DRAMless.

Phison’s entry-level NVMe solutions are missing from this comparison because their E8T controller is pretty old and unpopular at this point, and drives using the successor E13T and E19T controllers haven’t quite hit the market. The recently-released WD Blue SN550 adds the 1TB option that the earlier SN500 lacked. These two drives are particularly interesting because they are DRAMless NVMe SSDs that don’t use the Host Memory Buffer feature, but our test results from the 250GB SN500 don’t really provide a fair comparison against 1TB drives.

As always, it’s important to keep an eye on how entry-level NVMe drives perform relative to the more popular high-end models with 8-channel controllers and full-size DRAM caches, because the cheapest high-end drives are usually priced pretty close to entry-level NVMe SSDs. And lastly, representing the SATA SSD market segment, we include the Crucial MX500 as the mainstream choice, and the ADATA Ultimate SU750 featuring Realtek’s DRAMless 2-channel SATA controller.