The Windows 10 Fall Creators Update (FCU) has reportedly provided performance uplift under specific usage scenarios, most of which center around GPU-bound scenarios with Vega 56 or similar GPUs. We know with relative certainty that FCU has improved performance stability and frametime consistency with adaptive synchronization technologies – Gsync and FreeSync, mostly – and that there may be general GPU-bound performance uplift. Some of this could come down to driver hooks and implementation in Windows, some of it could be GPU or arch-specific. What we haven’t seen much of is CPU-bound tests, attempting to isolate the CPU as the DUT for benchmarking.

These tests look at AMD Ryzen R7 1700 (stock) performance in Windows 10 Creator’s Update (build 1703, ending in 608) versus Windows 10 Fall Creators Update. Our testing can only speak for our testing, as always, and we cannot reasonably draw conclusions across the hardware stack with these benchmarks. The tests are representative of the R7 1700 in CPU-bound scenarios, created by using a GTX 1080 Ti FTW3. Because this is a 1080 Ti FTW3, we have two additional considerations for possible performance uplift (neither of which will be represented herein):

  • - As an nVidia GPU, it is possible that driver/OS behavior will be different than with an AMD GPU
  • - As a 1080 Ti FTW3, it is possible and likely that GPU-bound performance – which we aren’t testing – would exhibit uplift where this testing does not

Our results are not conclusive of the entirety of FCU, and cannot be used to draw wide-reaching conclusions about multiple hardware configurations. Our objective is to start pinpointing performance uplift, and from what combination of components that uplift can be derived. Most reports we have seen have spotted uplift with 1070 or Vega 56 GPUs, which would indicate GPU-bound performance increases (particularly because said reports show bigger gains at higher resolutions). We also cannot yet speak to performance change on Intel CPUs.

Hardware news for the last week includes discussion on an inadvertent NZXT H700i case unveil (with “machine learning,” apparently), Ryzen/Vega APU, Vega partner card availability, and Coffee Lake availability.

Minor news items include the AMD AGESA 1.0.0.7 update to support Raven Ridge & Pinnacle Ridge, Noctua’s Chromax fans, and some VR news – like Oculus dropping its prices – and the Pimax 8K VR configuration.

Find the video and show notes below:

This episode of Ask GN was filmed a few days ago, but we ended up with so much content (like the H500P review and Vega 64 Strix PCB analysis) that we postponed its publication. The episode tackles popular topics of thermals and thermal testing, which have recently received more public interest, and also covers some top-level discussion of power, thermals, and electricity.

We spend most of the time discussing motherboard differences -- a story we've been harping on since January -- and how different board voltages affect CPUs in different ways. The rest of the intro is spent explaining thermal testing difficulties and challenges, and how we can best normalize for those in review content. The timestamps are below the video embed:

Following-up our tear-down of the ASUS ROG Strix Vega 64 graphics card, Buildzoid of Actually Hardcore Overclocking now visits the PCB for an in-depth VRM & PCB analysis. The big question was whether ASUS could reasonably outdo AMD's reference design, which is shockingly good for a card with such a bad cooler. "Reasonably," in this sentence, means "within reasonable cost" -- there's not much price-to-performance headroom with Vega, so any custom cards will have to keep MSRP as low as possible while still iterating on the cooler.

The PCB & VRM analysis is below, but we're still on hold for performance testing. As of right now, we are waiting on ASUS to finalize its VBIOS for best compatibility with AMD's drivers. It seems that there is some more discussion between AIB partners and AMD for this generation, which is introducing a bit of latency on launches. For now, here's the PCB analysis -- timestamps are on the left-side of the video:

We’re testing gaming while streaming on the R5 1500X & i5-8400 today, both CPUs that cost the same (MSRP is about $190) and appeal to similar markets. Difficulties stemming from stream benchmarking make it functionally impossible to standardize. CPU changes drastically impact performance during our streaming + gaming benchmarks, which means that each CPU test falls closer to a head-to-head than an overall benchmark. Moving between R5s and R7s, for instance, completely changes the settings required to produce a playable game + stream experience – and that’s good. That’s what we want. The fact that settings have to be tuned nearly on a per-tier basis means that we’re min-maxing what the CPUs can give us, and that’s what a user would do. Creating what is effectively a synthetic test is useful for outright component comparison, but loses resolution as a viable test candidate.

The trouble comes with lowering the bar: As lower-end CPUs are accommodated and tested for, higher-end components perform at lower-than-maximum throughput, but are capped in benchmark measurements. It is impossible, for example, to encode greater than 100% of frames to stream. That will always be a limitation. At this point, you either declare the CPU as functional for that type of encoding, or you constrict performance with heavier duty encoding workloads.

H264 ranges from Ultrafast to Slowest settings, with facets in between identified as Superfast, Veryfast, Faster, Fast, and Medium. As encoding speed approaches the Slow settings, quality enters into “placebo” territory. Quality at some point becomes indistinguishable from faster encoding settings, despite significantly more strain placed on the processor. The goal of the streamer is to achieve a constant framerate output – whether that’s 30FPS or 60FPS – while also maintaining a playable player-side framerate. We test both halves of the equation in our streaming benchmarks, looking at encode output and player output with equal discernment.

We’ve already sent off the information contained in this video to Buildzoid, who has produced a PCB & VRM analysis of the ROG Strix Vega 64 by ASUS. That content will go live within the next few days, and will talk about whether the Strix card manages to outmatch AMD’s already-excellent reference PCB design for Vega. Stay tuned for that.

In the meantime, the below is a discussion of the cooling solution and disassembly process for the ASUS ROG Strix Vega 64 card. For cooling, ASUS is using a similar triple-fan solution that we highly praised in its 1080 Ti Strix model (remarkable for its noise-normalized cooling performance), along with similar heatsink layout.

Learn more here:

We’re winding-down coverage of Vega, at this point, but we’ve got a couple more curiosities to explore. This content piece looks at a mix of clock scalability for Vega across a few key clocks (for core and HBM2), and hopes to constrain for a CU difference, to some extent. We obviously can’t fully control down to the shader level (as CUs carry more than just shaders), but we can get close to it. Note that the video content does generally refer to the V56 & V64 difference as one of shaders, but more than shaders are contained in the CUs.

In our initial AMD shader comparison between Vega 56 and Vega 64, we saw nearly identical performance between the cards when clock-matched to roughly 1580~1590MHz core and 945MHz HBM2. We’re now exploring performance across a range of frequency settings, from ~1400MHz core to ~1660MHz core, and from 800MHz HBM2 to ~1050MHz HBM2.

This content piece was originally written and filmed about ten days ago, ready to go live, but we then decided to put a hold on the content and update it. Against initial plans, we ended up flashing V64 VBIOS onto the V56 to give us more voltage headroom for HBM2 clocks, allowing us to get up to 1020MHz easily on V56. There might be room in there for a bit more of an OC, but 1020MHz proved stable on both our V64 and V56 cards, making it easy to test the two comparatively.

No surprise in that headline, really.

Some of this information is rehashed, but has been bulked-up by alleged AMD slides leaked to Informatica Cero. The slides, which are functionally in “rumor” status, indicate AMD’s code-named Matisse processors as launching in 2019. The Matisse CPUs will carry AMD’s Zen 2 architecture, but aim to continue supporting AM4 platforms. Before that launch, AMD’s Zen Plus iteration is targeted for 2018, allegedly, and will exist primarily as an optimization on the existing Ryzen CPUs. This can be thought of as an analog to the retired Intel tick-tock cadence, with Zen Plus likely targeting frequency tuning.

AMD’s architecture hasn’t generally shown a large gain from increasing CU count between top-tier and second-to-top cards. The Fury and Fury X, for instance, could be made to match with an overclock on the lower-tiered card. Additional gains on the higher-tiered card often amount from the increased power limit and clock, not from a straight shader increase. We’re putting that knowledge to the test on Vega architecture, equalizing the Vega 56 & Vega 64 clocks (and 945MHz HBM2 clocks) to determine how much of a difference emerges from the 4096 shaders on V64 to 3584 shaders on V56. Purely counting shaders, that’s a 14% increase to V64, but like most performance metrics, that won’t result in a linear performance increase.

We were able to crush Vega 64’s performance with our heavily modded Vega 56 card, using powerplay tables and liquid to jump to 1742MHz clock speeds. That's with modding, though, and isn't out-of-box performance -- it also doesn't give us any indication as to shader differences. Going less crazy about overclocking and limiting clocks to matched speeds, we can reveal the shader count difference.

It’s illegal to outright fix prices of products. Manufacturers have varying levels of sway when establishing cost to distributor partners and suggested retail prices, acted on much lower in the chain, and have to produce supply based on expectations of demand. We’ve previously talked about how MDF or other exchanges can be used to inspire retailers to work within some guidelines, but there are limits to the financial and legal extension of those means.

This context in mind, it makes sense that the undertone of discussion pertaining to video card prices – not just AMD’s, but nVidia’s – plants much of the blame squarely on retailers. There’s only so much that AMD and nVidia can do to drive prices at least somewhat close to MSRP. One of those actions is to put out more supply to sate demand but, as we saw during the last mining boom & bust (with emergent ASIC miners), there’s reason for manufacturers to remain hesitant of a major supply commitment. If AMD or nVidia were to place a large order with their fabs, there’d better be some level of confidence that the product will sell. Factory-to-shelf turn-around is a period of months, weeks of which can be shipping (unless opting for prohibitively expensive air freight).  A period of months is a wide window. We’ve seen mining markets “crash” and recover in a period of days, or hours, with oft unpredictable frequency and intensity. That’d explain why AMD might be hesitant to issue large orders of older product, like the RX 500 series, to try and meet demand.

Page 1 of 32

We moderate comments on a ~24~48 hour cycle. There will be some delay after submitting a comment.

Advertisement:

  VigLink badge