HTPC Aspects : Miscellaneous Factors

In this section, we cover some miscellaneous HTPC aspects that are too short to warrant a separate section. These include a discussion of various display refresh rates supported, a short look at the hardware encoder (NVENC) in action and a summary of our thoughts on the GT 750Ti as a HTPC GPU.

Refresh Rate Accuracy:

NVIDIA provides an easy way to customize refresh rates. The process remains the same as what we explained in our review of the GT 640. The 23 Hz setting gives us a refresh rate of 23.971 Hz. With Intel providing rock-solid 23.976 Hz support in Haswell, it is time NVIDIA got the out-of-the-box refresh rate support correct.

NVIDIA also allows setting of refresh rates not reported as available by the display's EDID. On the Sony KDL46EX720, it allowed driving of 1080p50 without any issues. The flexibility is definitely appreciated, though it would be nice to have better accuracy without all the tweaking.

Hardware Encoder: NVENC

We used CyberLink MediaEspresso v6.7 to evaluate the hardware encoder block. Our test clip was a 3-minute long 1080p24 H.264 stream at 36 Mbps and the target was a 720p24 H.264 stream at 6 Mbps. The time taken for conversion and the power consumption at the wall during the conversion process are provided in the table below.

GPU Video Encoding Performance
  Conversion Time Power
NVIDIA GeForce GTX 750 Ti 2:54 88.97W
NVIDIA GeForce GT 640 0:36 108.18W
AMD Radeon HD 7750 (VCE) 1:06 76.84W
Intel HD 4000 QuickSync (Better Quality/Fast Conversion) 0:24 63.91W

It appears as if the 750Ti is using the CUDA path rather than NVENC, while the 640 seems to use NVENC fine. We had readied ourselves for some quality comparison using objective metrics for the new NVENC. It looks like we have to wait for this issue to be resolved before proceeding down that path. [Update: NVIDIA got back to us indicating that this is a Maxwell-related driver issue. We are waiting for new drivers]

HTPC Verdict - Wait and Watch

We have taken a look at the HTPC credentials of the 750Ti and compared it with the GT 640 and the HD 7750. In terms of power efficiency, it is hard not to recommend the 750Ti. With a 60W TDP, it is amenable to passive cooling also. However, it comes to the market at a time when the HEVC standard has just been ratified (preventing it from having a full-blown hardware accelerated decoder) and HDMI 2.0 with 4Kp60 support being right around the corner. The perfect HTPC GPU would include support for both, but the 750Ti, unfortunately, is a bit early to the game. More troublesome is the fact that CyberLink's MediaEspresso seems unable to take advantage of the new NVENC and the fact that some of our 1080p60 H.264 clips are showing decoding artifacts (considering they play perfectly using the GT 640).

We would suggest HTPC enthusiasts to adopt a wait-and-watch approach to the GT 750Ti, particularly with respect to driver bugs specific to the 750Ti and also the extent of HEVC decode support that will be available. Depending on the requirements, it might also be prudent to wait for a Maxwell GPU with HDMI 2.0 support.

HTPC Aspects : Decoding & Rendering Benchmarks The Test
Comments Locked

177 Comments

View All Comments

  • Harry Lloyd - Tuesday, February 18, 2014 - link

    20 nm Maxwell will be epic. Gimme.
  • TheinsanegamerN - Tuesday, February 18, 2014 - link

    Imagine. OCed Geforce 690 level performance, out of a single chip, with 8 GB of RAM on a 512 bit bus, pulling the same amount of power as a geforce 770. One can dream....
  • ddriver - Tuesday, February 18, 2014 - link

    LOL, epic? Crippling FP64 performance further from 1/24 to 1/32 - looks like yet another nvidia architecture I'll be skipping due to abysmal compute performance per $ ratio...
  • JDG1980 - Tuesday, February 18, 2014 - link

    This card is designed for gaming and HTPC. Only a tiny fraction of users need FP64.
  • nathanddrews - Tuesday, February 18, 2014 - link

    So I guess we'll have to wait for the 750TIB before we can see SLI benchmarks. Two of these would be within reach of 770 while using considerably less power. Hypothetically, that is.
  • ddriver - Tuesday, February 18, 2014 - link

    You do realize the high end GPUs on the same architecture will have the same limitation?
  • Morawka - Tuesday, February 18, 2014 - link

    I thought the higher end Maxwell cards will have Denver/aRM cores on the PCB as well.
  • Mr Perfect - Wednesday, February 19, 2014 - link

    It might be a software/firmware limitation though. From what the compute enthusiasts have said, the only difference between the Titan's full compute and 780Ti's cut down compute is firmware based. They've got the same chip underneath, and some people hack their 780s for full compute. They're probably doing the same thing with the Maxwell stack.
  • chrnochime - Wednesday, February 19, 2014 - link

    Got link for the hack? Sounds interesting.
  • Mr Perfect - Thursday, February 20, 2014 - link

    I don't myself, but if you're interested look up IvanIvanovich over at bit-tech.net. He was talking about vbios mods and resistor replacement tweaks that can do that.

Log in

Don't have an account? Sign up now