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

Forums - Nintendo Discussion - Clearing up a major misconception about PowerPC

mine on 30 January 2015
Captain_Tom said:
walsufnir said:


Far easier to program for? What? Especially arm... no, the power isa is known since many years and IBM is known for especially good compilers.

At the very least it is better known by gaming programmers, and thus easier.  All major game engines are built to work on them, making a port to Wii U annoying.

WHAT is known by the gaming programmers? 

Its the GAME ENGINES.

So what you are saying is that XOne and PS4 get a higher percentage of mediocre games which won't be able to utilize the consoles to their maximum as those game engines are built to run on ALL supported platforms. 

Of course it is ALWAYS easier AND FASTER to developer MEDIOCRE games which have problems rendering 60fps on hardware which is more than capable of delivering it... .

"Gamers" don't care and they can be easily fooled with "common sense" and "better known" stuff... 



Around the Network
fatslob-:O on 30 January 2015
mine said:


Not common sense. Common sense is that comilers do the hard work for the programmer. The most important thing the developer can do is studying the effects of their data structures, data and program flow on the CPU caches and act accordingly.

So: as the Wii U has MORE L2 cache per Core than the XOne and PS4 Cores. Code runs on the Wii U cores more efficient. But the HD twins make it more then up with more cores and - of course - a better GPU.

The HD twins still have more CPU cache ...

BTW: take a look at the PS4 developer presentations. The small L2 cache is hurting BIG time when CPU AND GPU are accessing the GDDR...

Link ?

Or the other way round: the Wii Us big L2 cache, big eDRAM and balanced CPU core / GPU enabled to deliver more than most people expected from such a configuration...



nomad on 30 January 2015
Captain_Tom said:
fatslob-:O said:

How does that factor into making it easier to program for those devices when a compiler could just simply generate instructions for multiple ISAs ?


LOL it's not just that "Simple."  I am only saying what has been repeated by numerous devs...and common sense of course.

Game devs and software devs have been working on PPC for years, porting PS3/X360 to PC, x86 to POWER servers, PPC to ARM, etc. There's a lot of documentation and toolchains available for these kind of work. The reason why porting to WiiU is "difficult", is because it is simply too weak relative to PS4/X1, but the main reason Publishers don't want to make games for WiiU is because they don't think there games can sell enough to make money. PPC has little to do with it. Most game programmers these days now code in a high level langauge anyway, C++(high-low depending on who you ask),Java, C#, etc. is the same langauge on an ARM, x86, or PPC device. They just have to keep in mind the different standards/conventions and capabilities of each cpu and device. There may be some assembly and compiler code that neededs to be done to improve performance, but doing it different on one does not neccesarilly make programming more difficult than on the other.

From what I see, the annonymous compaints about difficulties working on WiiU was mainly during the early stage of WiiU, when devs with early dev kits got little support from Nintendo. And their complaints about the CPU was about its power/performance. Keep in mind, DarksiderII devs got their engine working very quick with just a few guys. There are other report about Nintendo's online infrastructure but those have little to do with PPC and WiiU's hardware.



curl-6 on 30 January 2015
walsufnir said:
captain carot said:

As for Espresso, it likely isnt a 970. Its as well no stock 750 PPC. THere where already changes made for Gekko, like some SIMD, FPU...

I go with marcan in this context as he is quite a knowledgable guy when it comes to Wii/WiiU hardware and many things we know about the Espresso is coming from him and his findings. In general it's of course not a stock 750 but derived from it and I guess if he says it's 750 we can likely believe him in this context. 

 

Edit: Concerning the ISA I found this: http://www.radgametools.com/bnkhist.htm

"Added Wii-U support for Bink 2 - play 30 Hz 1080p or 60 Hz 720p video! We didn't think this would be possible - the little non-SIMD CPU that could!" I didn't know it doesn't feature SIMD...

According to Marcan it has SIMD, just weaker SIMD compared to its competitors.

They basically compromised raw power in exchange for backards compatibility. That said, it has some advantages over the last gen HD twins that help mitigate its lower clock speed; better IPC, out of order execution, the ability to offload work to the GPGPU and dedicated audio chip, etc.



Marcusius on 30 January 2015
WolfpackN64 said:

I know this is in the Nintendo Discussion, but that's because they still use PowerPC chips in their console.

Nearly everytime a topic handles the internals of the Wii U, some comments talk about the "dated PowerPC architecture". This is complete bollocks. The Wii U chipset is derived from a dated PowerPC chip (Espresso, derived from Broadwell, derived from Gecko, derived from PowerPC 750), that is correct, but the PowerPC Architecture itself is NOT dated.

The PowerPC architecture saw the light of day in 1992 with the current form Power ISA v2.07 released in 2013. For comparison, x86 started in 1978 with the most recent implementation x86-64 or AMD64 being 2003.

There, needed to get that of my chest.

Another misbeleive is the PowerPC is less powerfull than an X86.

Wrong, the CPU cost more but the CPU is much more efficient. 

PowerPC is a RISC meaning it need less cpu cycles to do the same task.



Around the Network
fatslob-:O on 30 January 2015
curl-6 said:

According to Marcan it has SIMD, just weaker SIMD compared to its competitors.

They basically compromised raw power in exchange for backards compatibility. That said, it has some advantages over the last gen HD twins that help mitigate its lower clock speed; better IPC, out of order execution, more cache, the ability to offload work to the GPGPU and dedicated audio chip, etc.

It's SIMD capabilities is limited to 2 single precision floats! (According to the source section 3.4)

It doesn't even extend to integers! SSE and SSE2 which are extensions practically found on EVERY modern x86 CPUs ever since the PENTIUM 4 can compute 4 single precision floats or 8 16-bit integers ...



curl-6 on 30 January 2015
fatslob-:O said:
curl-6 said:

According to Marcan it has SIMD, just weaker SIMD compared to its competitors.

They basically compromised raw power in exchange for backards compatibility. That said, it has some advantages over the last gen HD twins that help mitigate its lower clock speed; better IPC, out of order execution, more cache, the ability to offload work to the GPGPU and dedicated audio chip, etc.

It's SIMD capabilities is limited to 2 single precision floats! (According to the source section 3.4)

It doesn't even extend to integers! SSE and SSE2 which are extensions practically found on EVERY modern x86 CPUs ever since the PENTIUM 4 can compute 4 single precision floats or 8 16-bit integers ...

It has a 32-bit integer unit.



fatslob-:O on 30 January 2015
curl-6 said:

It has a 32-bit integer unit.

Yes I know that but the integer unit still only executes 1 element at a time ...



curl-6 on 30 January 2015
fatslob-:O said:
curl-6 said:

It has a 32-bit integer unit.

Yes I know that but the integer unit still only executes 1 element at a time ...

Better than not having one though.



fatslob-:O on 30 January 2015
curl-6 said:

Better than not having one though.

Give me a list of instruction set architectures which can't use integers ...