Moore Thread's MTT S80, World's First PCIe Gen 5 Gaming Graphics Card, Now Priced at $164

[btarunr]

The Moore Thread's MTT S80 discrete graphics card is now available as part of a special 11-11 (Single's Day) promotion in China, for the equivalent of USD $164, making it both the world's first PCIe Gen 5 gaming graphics card, and the most affordable one to feature 16 GB of memory. Moore Thread's is a Chinese GPU manufacturer that has been aiming to build a contemporary GPU to grab a slice of the entry-mainstream gaming market in China for a few years now.

Much of the PC gaming scene in China doesn't involve AAA productions in need of the fastest GPU out there, but rather GPUs from the mainstream performance tier—Moore Thread's knows this, and has been reinventing many wheels in the absence of the kind of graphics IP cross-licensing entanglement that exists among NVIDIA, AMD, and Intel. The company's fastest GPU is the MTT S80 launched in late 2022, which has the bragging rights to be the world's first with a PCI Express Gen 5 bus interface. Does it need this kind of bandwidth? We honestly don't know, after seeing how sensitive to PCIe interface and resizable-BAR even mainstream Intel GPUs can be. At launch the performance level of the MTT S80 made it more of a novelty than anything, with performance barely matching a Radeon RX 6400, making it about as fast as the iGPU of AMD's Ryzen 5000G "Cezanne" desktop APUs. This is just enough for China's homebrew MOBAs and MMORPGs that are designed to maximize market reach, and hence tend to contain a lot of pre-baked content.



Image Courtesy: Expreview

View at TechPowerUp Main Site | Source

 

[natr0n]

I wonder whose drivers they stole to get this going.

 

[Deleted member 185088]

That's a step in the right direction, our hope is China to have more competition in the market, though it might take some time to materialise.

I wonder whose drivers they stole to get this going.

They built their own.

 

[tabascosauz]

Isn't this the same power-hungry one that looked like it had decent compute performance but had absolutely hilarious game performance with woefully unprepared drivers? PowerVR based iirc

They built their own

Unless Moore Threads has been busting their asses for the past half year furiously accomplishing Intel-levels of driver catch-up work........yeah, no. Apart from some lukewarm news buzz in September about a driver update for OpenGL optimizations, seems like Moore Threads hasn't budged on the driver front since everyone did their testing of the S80. I keep seeing this 40% improvement being thrown around, but what is +40% of "basically doesn't run"?

 

[sam_86314]

I thought I had heard these were based on Imagination's PowerVR core.

Used to be a fairly common GPU architecture in phones, and even dates back to being used in the Dreamcast.

 

[TumbleGeorge]

With one of latest drivers this card is faster than GTX 1650. Much better than first time were benchmarked.

 

[AusWolf]

I'll just drop this here:

 

[TumbleGeorge]

I'll just drop this here:

It's old from July, 15th.

 

[AusWolf]

It's old from July, 15th.

I know - just in case anyone is curious about the card itself.

 

[tabascosauz]

With one of latest drivers this card is faster than GTX 1650. Much better than first time were benchmarked.

My guy:

0.8x render resolution, all low settings, FSR2 AA 1080p and it can't keep 60fps in simple fights and Liyue main street. That's not just slightly worse than a Vega 7 iGPU; the only things that bog down Vega below 60fps are some night city scenes, complex battles, and some Sumeru/Fontaine scenes, and all at 1.0x render resolution and a mix of medium/low settings because shadows settings hit iGPUs hard.

Like damn, Genshin is definitely not the greatest GPU benchmark ever, but this isn't any less trash than we were already seeing from S80. If you just based your statement on the "198,000 points" in the mystery unnamed benchmark, then keep coping I guess.

 

[AusWolf]

My guy:

View attachment 319101

0.8x render resolution, all low settings, FSR2 AA 1080p and it can't keep 60fps in simple fights and Liyue main street. That's not just slightly worse than a Vega 7 iGPU; the only things that bog down Vega below 60fps are some night city scenes, complex battles, and some Sumeru/Fontaine scenes, and all at 1.0x render resolution and a mix of medium/low settings.

Like damn, Genshin is definitely not the greatest GPU benchmark ever, but this isn't any less trash than we were already seeing from S80. If you just based your statement on the "198,000 points" in the mystery benchmark, then keep coping I guess.

Yeah, but does the Vega 7 iGPU consume 110+ Watts at idle? Moore Threads rocks!

 

[Deleted member 185088]

Unless Moore Threads has been busting their asses for the past half year furiously accomplishing Intel-levels of driver catch-up work........yeah, no. Apart from some lukewarm news buzz in September about a driver update for OpenGL optimizations, seems like Moore Threads hasn't budged on the driver front since everyone did their testing of the S80. I keep seeing this 40% improvement being thrown around, but what is +40% of "basically doesn't run"?

Intel is a perfect example as despite making GPUs forever ARC couldn't run DX9 games natively, one should limit their expectations.

+X% is marketing rubbish, Apple newest SOC is a good example where it went backwards in efficiency.

 

[tabascosauz]

Yeah, but does the Vega 7 iGPU consume 110+ Watts at idle? Moore Threads rocks!

It's all about More right? More X's in the name, more W's at idle (who doesn't want a free W?), even more mores - even more has an extra O!

Correction to my statement: it is significantly worse than Vega 8, but the picture the September testers portray is roughly where a default clocks Vega 7 is in Genshin, if not worse. Which is to say, trash, because it performs like an iGPU (synthetic benchmark queen) without being one. So I guess if the original drivers couldn't even run the game at all, then it's still an infinite performance uplift?

I don't even see any actual results showing it as faster than 1650 in the linked "review". GTX 1650 desktop is quite a bit faster than mobile 680M and GTX 1650 mobile is slightly faster than 680M, while 680M with slow DDR5 breezes through Genshin at 1080p high settings without missing a beat. 680M in turn trashes Vega 8 even with dialed up core and fast RAM, and Vega 8 in turn demolishes whatever the S80 seems to be doing.

 

[TumbleGeorge]

then keep coping I guess.

He promises more tests. So we'll wait for him to do them.

 

[AusWolf]

It's all about More right? More X's in the name, more W's at idle (who doesn't want a free W?), even more mores - even more has an extra O!

Whether it's frames per second, Watts, number of fans, calories... a proper gamer doesn't care. A proper gamer only wants BIGGER NUMBERS!

 

[P4-630]

Just lovely.......

 

[chrcoluk]

Is it capped to 4 lanes? 5x4 is worse than 3x16 as it can only run well on newest boards. Or is it actually 5x16.

 

[AusWolf]

Is it capped to 4 lanes?

It is capped to being shit, as far as I saw in the review.

 

[BlaezaLite]

It is capped to being shit, as far as I saw in the review.

I was going to say give them a chance to mature and grow, but no, it's rabid, shoot it.

 

[Tropick]

"for the equivalent of USD $164"

Yeah best I can do is tree fiddy.

 

[TumbleGeorge]

"for the equivalent of USD $164"

Yeah best I can do is tree fiddy.

When you do, post some benchmarks and game results here, or in an appropriate forum thread, please.

 

[unwind-protect]

The drivers are always the problem.

And it shouldn't be necessary to support specific games one by one. A conformant implementation of the API should make all games work.

This is like an Ethernet chip driver that only supports Chrome but you have to wait for Firefox "support".

 

[Tropick]

When you do, post some benchmarks and game results here, or in an appropriate forum thread, please.

If I was actually able to get my hands on the thing I would love to do testing.

 

[Vayra86]

He promises moore tests.

ftfy

This is the funniest GPU I've seen in years tbh

 

[AnotherReader]

It's clear that the blame for the lackluster performance lies with Moore Thread's driver team. The GPU is reputed to be based on Imagination's IP which is also used by Apple and we have seen that Apple's GPUs are much faster than this.