TU117: Tiny Turing

Before we take a look at the Zotac card and our benchmark results, let’s take a moment to go over the heart of the GTX 1650: the TU117 GPU.

TU117 is for most practical purposes a smaller version of the TU116 GPU, retaining the same core Turing feature set, but with fewer resources all around. Altogether, coming from the TU116 NVIDIA has shaved off one-third of the CUDA cores, one-third of the memory channels, and one-third of the ROPs, leaving a GPU that’s smaller and easier to manufacture for this low-margin market.

Still, at 200mm2 in size and housing 4.7B transistors, TU117 is by no means a simple chip. In fact, it’s exactly the same die size as GP106 – the GPU at the heart of the GeForce GTX 1060 series – so that should give you an idea of how performance and transistor counts have (slowly) cascaded down to cheaper products over the last few years.

Overall, NVIDIA’s first outing with their new GPU is an interesting one. Looking at the specs of the GTX 1650 and how NVIDIA has opted to price the card, it’s clear that NVIDIA is holding back a bit. Normally the company launches two low-end cards at the same time – a card based on a fully-enabled GPU and a cut-down card – which they haven’t done this time. This means that NVIDiA is sitting on the option of rolling out a fully-enabled TU117 card in the future if they want to.

By the numbers, the actual CUDA core count differences between GTX 1650 and a theoretical fully-enabled GTX 1650 Ti are quite limited – to the point where I doubt a few more CUDA cores alone would be worth it – however NVIDIA also has another ace up its sleeve in the form of GDDR6 memory. If the conceptually similar GTX 1660 Ti is anything to go by, a fully-enabled TU117 card with a small bump in clockspeeds and 4GB of GDDR6 could probably pull far enough ahead of the vanilla GTX 1650 to justify a new card, perhaps at $179 or so to fill NVIDIA’s current product stack gap.

The bigger question is where performance would land, and if it would be fast enough to completely fend off the Radeon RX 570. Despite the improvements over the years, bandwidth limitations are a constant challenge for GPU designers, and NVIDIA’s low-end cards have been especially boxed in. Coming straight off of standard GDDR5, the bump to GDDR6 could very well put some pep into TU117’s step. But the price sensitivity of this market (and NVIDIA’s own margin goals) means that it may be a while until we see such a card; GDDR6 memory still fetches a price premium, and I expect that NVIDIA would like to see this come down first before rolling out a GDDR6-equipped TU117 card.

Turing’s Graphics Architecture Meets Volta’s Video Encoder

While TU117 is a pure Turing chip as far as its core graphics and compute architecture is concerned, NVIDIA’s official specification tables highlight an interesting and unexpected divergence in related features. As it turns out, TU117 has incorporated an older version of NVIDIA’s NVENC video encoder block than the other Turing cards. Rather than using the Turing block, it uses the video encoding block from Volta.

But just what does the Turing NVENC block offer that Volta’s does not? As it turns out, it’s just a single feature: HEVC B-frame support.

While it wasn’t previously called out by NVIDIA in any of their Turing documentation, the NVENC block that shipped with the other Turing cards added support for B(idirectional) Frames when doing HEVC encoding. B-frames, in a nutshell, are a type of advanced frame predication for modern video codecs. Notably, B-frames incorporate information about both the frame before them and the frame after them, allowing for greater space savings versus simpler uni-directional P-frames.

I, P, and B-Frames (Petteri Aimonen / PD)

This bidirectional nature is what make B-frames so complex, and this especially goes for video encoding. As a result, while NVIDIA has supported hardware HEVC encoding for a few generations now, it’s only with Turing that they added B-frame support for that codec. The net result is that relative to Volta (and Pascal), Turing’s NVENC block can achieve similar image quality with lower bitrates, or conversely, higher image quality at the same bitrate. This is where a lot of NVIDIA’s previously touted “25% bitrate savings” for Turing come from.

Past that, however, the Volta and Turing NVENC blocks are functionally identical. Both support the same resolutions and color depths, the same codecs, etc, so while TU117 misses out on some quality/bitrate optimizations, it isn’t completely left behind. Total encoder throughput is a bit less clear, though; NVIDIA’s overall NVENC throughput has slowly ratcheted up over the generations, in particular so that their GPUs can serve up an ever-larger number of streams when being used in datacenters.

Overall this is an odd difference to bake into a GPU when the other 4 members of the Turing family all use the newer encoder, and I did reach out to NVIDIA looking for an explanation for why they regressed on the video encoder block. The answer, as it turns out, came down to die size: NVIDIA’s engineers opted to use the older encoder to keep the size of the already decently-sized 200mm2 chip from growing even larger. Unfortunately NVIDIA isn’t saying just how much larger Turing’s NVENC block is, so it’s impossible to say just how much die space this move saved. However, that the difference is apparently enough to materially impact the die size of TU117 makes me suspect it’s bigger than we normally give it credit for.

In any case, the impact to GTX 1650 will depend on the use case. HTPC users should be fine as this is solely about encoding and not decoding, so the GTX 1650 is as good for that as any other Turing card. And even in the case of game streaming/broadcasting, this is (still) mostly H.264 for compatibility and licensing reasons. But if you fall into a niche area where you’re doing GPU-accelerated HEVC encoding on a consumer card, then this is a notable difference that may make the GTX 1650 less appealing than the TU116-powered GTX 1660.

The NVIDIA GeForce GTX 1650 Review: Featuring ZOTAC Meet the ZOTAC GeForce GTX 1650 OC


View All Comments

  • dromoxen - Monday, May 20, 2019 - link

    Its the current person mopping the floor who designed AMD's last generation of gfx cards.
    Another reason to buy this Crda is that you may not want the heat produced . I for one have started to use a 10w NUC in prefernece to a 75w HTPC just becuase the heating effect is less . UK,not jamaica or Saudi
  • plonk420 - Friday, May 3, 2019 - link

    thanks for all the compute benches! yuuuugely appreciated! Reply
  • ads295 - Friday, May 3, 2019 - link

    Can I use this to play ten year old games in full glory at 1440p? Reply
  • Ryan Smith - Friday, May 3, 2019 - link

    Easily. Heck, depending on the game, you could probably get away with doing that on an iGPU. Reply
  • Ashinjuka - Saturday, May 4, 2019 - link

    Probably not full-glory S.T.A.L.K.E.R. Definitely not full-glory S.T.A.L.K.E.R. with graphics mods. Reply
  • SaturnusDK - Friday, May 3, 2019 - link

    Quite frankly at the $150, no one, and I do mean no one should buy this card. Even if you refurb an old OEM system the price difference up to an RX570 lets you buy a decent 80+ certified power supply and have a system that is more powerful and probably more power efficient at the same time. A standard OEM PSU in a an old computer is so inefficient that just replacing it makes up for more than the power consumption difference between a 1650 and an RX570. And gives you at least 15% more performance for the same amount of money spent. Reply
  • Oxford Guy - Saturday, May 4, 2019 - link

    I doubt anyone should have purchased the 960 and yet it's the 5th most popular Steam card.

    This place didn't even bother to review it.
  • RSAUser - Friday, May 3, 2019 - link

    A 1060 costs the same price as this 1650 here, I see no reason to buy it. Terrible value for money. Reply
  • RSAUser - Friday, May 3, 2019 - link

    You can't compare the 1650 to the 950, they're priced completely differently at launch. Stop going directly with the product number. The 1650 is between 960 and 970. Reply
  • linuxgeex - Friday, May 3, 2019 - link

    "Notably, B-frames incorporate information about both the frame before them and the frame after them, allowing for greater space savings versus simpler uni-directional P-frames."

    No. H.264 and H.265 (AVC/HEVC) have (optional) bi-directional P-Frames. That increases the complexity of the search required to create a B-Frame which would use significantly less data than a P-Frame. A lower-capability GPU may not be able to perform that search in real time, and in that case there's no point implementing it, even if it would increase compression efficiency, because the selling point of hardware HEVC compression is that it can be done in real time.
    B-Frames are simpler than P-Frames. Not the other way around.
    To be clear: I-Frames are effectively a still shot of the scene, like a JPEG.
    P-Frames hold motion data with references to I-Frames and P-Frames - they encode linear motion for blocks in the image, they encode replacement blocks for new data needed to replace changes, ie when something moves over a background and reveals what was behind it.
    If B-Frames are used, then intermediate frames are calculated between the P-Frames and their references based on their encoded block motion data. These result in what are called "tweens" in animation - images that are partway between a start and an end. The B-Frames encode small fixes for errors in the guessed (by linear interpolation) intermediate frames. The less motion there is, and the more linear the motion is, the more accurate the interpolated frames are and the more B-Frames you can have between P-Frames before the B-Frames become necessarily larger than a new P-Frame would have been. Generating those B-Frames and estimating / discarding them based on whether they can be as efficient as the P-Frames is a lot of work even when the P-Frames don't have bidirectional references. HEVC allows for more than just bidirectional (2 frame) motion prediction references. It allows using an P-Frame to inherit any other P-Frame's motion references and it allows P-Frames to target a B-Frame for motion estimation. That introduces an order of magnitude more search possibilities than H.264/AVC. HEVC with B-Frames disabled basically performs at a similar efficiency to AVC because all those options are off the table.

Log in

Don't have an account? Sign up now