By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Nintendo Discussion - Clarifying the 1.5TFLOPS of the SWITCH for those who just see the numbers.

FunFan on 02 November 2016
Zkuq said:
I'm no expert on graphics or game programming, but I'm under the impression that there's quite a limited range of situations where half-precision FP numbers are usable. Any experts capable of shedding some light on this? How much of a boost does FP16 give over FP32 without too much noticeable degradation in image quality (or other areas) in a typical modern game?

It seems so, but 32-bit has been the standar for a long time, so its probably strongly established in devs mindset. Only time will tell if theres any advantage to 16-bit flops, if developers even take their time to experiment with it.



“Simple minds have always confused great honesty with great rudeness.” - Sherlock Holmes, Elementary (2013).

"Did you guys expected some actual rational fact-based reasoning? ...you should already know I'm all about BS and fraudulence." - FunFan, VGchartz (2016)

Around the Network
Zkuq on 02 November 2016
FunFan said:
Zkuq said:
I'm no expert on graphics or game programming, but I'm under the impression that there's quite a limited range of situations where half-precision FP numbers are usable. Any experts capable of shedding some light on this? How much of a boost does FP16 give over FP32 without too much noticeable degradation in image quality (or other areas) in a typical modern game?

It seems so, but 32-bit has been the standar for a long time, so its probably strongly established in devs mindset. Only time will tell if theres any advantage to 16-bit flops, if developers even take their time to experiment with it.

I don't think it's just that we've been using FP32 for so long. To give you an idea about how imprecise FP16 is: the maximum integer it can accurately represent is 65504. It gets worse with decimal numbers: you can't even represent anything larger than 1024 as a decimal number anymore because FP16 simply lacks the precision for that. Obviously this gets worse when you perform operations with them. Wikipedia has some more details on the precision of FP16. This is actually a good reminder for me that the further away you move from 0 with FP numbers, the less precision you have.



FunFan on 02 November 2016
Zkuq said:
FunFan said:

It seems so, but 32-bit has been the standar for a long time, so its probably strongly established in devs mindset. Only time will tell if theres any advantage to 16-bit flops, if developers even take their time to experiment with it.

I don't think it's just that we've been using FP32 for so long. To give you an idea about how imprecise FP16 is: the maximum integer it can accurately represent is 65504. It gets worse with decimal numbers: you can't even represent anything larger than 1024 as a decimal number anymore because FP16 simply lacks the precision for that. Obviously this gets worse when you perform operations with them. Wikipedia has some more details on the precision of FP16. This is actually a good reminder for me that the further away you move from 0 with FP numbers, the less precision you have.

I already know that, and I'm still stickinf to what I said: "Only time will tell if theres any advantage to 16-bit flops, if developers even take their time to experiment with it."



“Simple minds have always confused great honesty with great rudeness.” - Sherlock Holmes, Elementary (2013).

"Did you guys expected some actual rational fact-based reasoning? ...you should already know I'm all about BS and fraudulence." - FunFan, VGchartz (2016)

Zkuq on 02 November 2016
FunFan said:
Zkuq said:

I don't think it's just that we've been using FP32 for so long. To give you an idea about how imprecise FP16 is: the maximum integer it can accurately represent is 65504. It gets worse with decimal numbers: you can't even represent anything larger than 1024 as a decimal number anymore because FP16 simply lacks the precision for that. Obviously this gets worse when you perform operations with them. Wikipedia has some more details on the precision of FP16. This is actually a good reminder for me that the further away you move from 0 with FP numbers, the less precision you have.

I already know that, and I'm still stickinf to what I said: "Only time will tell if theres any advantage to 16-bit flops, if developers even take their time to experiment with it."

Oh, I see. I didn't have a problem with your point anyway, it's just that one detail I found a bit dubious: :P But yeah, I agree with you.



FunFan on 02 November 2016
Zkuq said:
FunFan said:

I already know that, and I'm still stickinf to what I said: "Only time will tell if theres any advantage to 16-bit flops, if developers even take their time to experiment with it."

Oh, I see. I didn't have a problem with your point anyway, it's just that one detail I found a bit dubious: :P But yeah, I agree with you.

Of course it was dubious, I was just making an out-of-my-ass assumption. But I kind of do that all the time. Guess you didn't noticed.



“Simple minds have always confused great honesty with great rudeness.” - Sherlock Holmes, Elementary (2013).

"Did you guys expected some actual rational fact-based reasoning? ...you should already know I'm all about BS and fraudulence." - FunFan, VGchartz (2016)

Around the Network
JEMC on 02 November 2016
vivster said:
fleischr said:

That assumes the Parker chip in Switch is vanilla X2. But being a "custom" Tegra it could be more powerful. Or less powerful. I hope nVidia distinguishes it as something like "Tegra N" rather than "Tegra X1"/ "Tegra X2"

I don't think nVidia and Nintendo would both decide to use the exact same chip in cars in the Switch. A gaming device has more use of extra GPU than a car.

Different use-cases. Different products.

I know Tegra is mostly used for cars with smartphones and tablets being more of a fringe case but that doesn't mean it's just made for it. The X2 (or P1) has the most powerful Tegra graphics yet. Honestly I would be surprised If Nintendo's device will have anything above the specifications beside a higher clock in dock mode.

The Tegra X2 is very CPU centric, featuring the same 256 shaders of the X1 but updating them from Maxwell to Pascal, but increasing the CPU cores from 4xA57+4xA53 cores in a LITTLE.big configuration to 2xDenver and 4xA57 cores in a SUPER.big config.

I think one of the customisations to turn the X2 into a more capable gaming chip would be getting rid of some of those CPU cores and add more shaders.



Please excuse my bad English.

Currently gaming on a PC with an i5-4670k@stock (for now), 16Gb RAM 1600 MHz and a GTX 1070

Steam / Live / NNID : jonxiquet    Add me if you want, but I'm a single player gamer.

vivster on 03 November 2016
JEMC said:
vivster said:

I know Tegra is mostly used for cars with smartphones and tablets being more of a fringe case but that doesn't mean it's just made for it. The X2 (or P1) has the most powerful Tegra graphics yet. Honestly I would be surprised If Nintendo's device will have anything above the specifications beside a higher clock in dock mode.

The Tegra X2 is very CPU centric, featuring the same 256 shaders of the X1 but updating them from Maxwell to Pascal, but increasing the CPU cores from 4xA57+4xA53 cores in a LITTLE.big configuration to 2xDenver and 4xA57 cores in a SUPER.big config.

I think one of the customisations to turn the X2 into a more capable gaming chip would be getting rid of some of those CPU cores and add more shaders.

Then they might as well just go with an X1 shrink.



If you demand respect or gratitude for your volunteer work, you're doing volunteering wrong.

walsufnir on 03 November 2016
Zkuq said:
I'm no expert on graphics or game programming, but I'm under the impression that there's quite a limited range of situations where half-precision FP numbers are usable. Any experts capable of shedding some light on this? How much of a boost does FP16 give over FP32 without too much noticeable degradation in image quality (or other areas) in a typical modern game?

It isn't really faster just because you lose precision. It *could* be faster if

a) the algorithm or the computation doesn't need to be fp32 in all cases and

b) the hardware can store and use two 16 bit floating point numbers in the same time it would store and use one fp32.

Everyone is free to use fp16 right now but there is no real benefit for that.



vivster on 03 November 2016
walsufnir said:
Zkuq said:
I'm no expert on graphics or game programming, but I'm under the impression that there's quite a limited range of situations where half-precision FP numbers are usable. Any experts capable of shedding some light on this? How much of a boost does FP16 give over FP32 without too much noticeable degradation in image quality (or other areas) in a typical modern game?

It isn't really faster just because you lose precision. It *could* be faster if

a) the algorithm or the computation doesn't need to be fp32 in all cases and

b) the hardware can store and use two 16 bit floating point numbers in the same time it would store and use one fp32.

Everyone is free to use fp16 right now but there is no real benefit for that.

Shouldn't it be easy to implement though? Developers mark the stuff that is FP16 and the GPU driver does the rest.



If you demand respect or gratitude for your volunteer work, you're doing volunteering wrong.

Soundwave on 03 November 2016

FP16 is largely useless for games. I don't see many developers really using it, not even Nintendo.

Nintendo is about pre-baked hardware, since the N64 days of nightmarish microcode, they hate really digging into a chip and coding to the metal.

The FP32 number is the real performance you're going to get with video games.