The basic point of your post is accurate, bluewings is an idiot, but there is virtually zero information available that would support this:

 

Virtually nothing has been made public about he F-22's EW suite

 

Well, except for Keltec bragging about the F-22 using its tubes in its EW suite...

 

Especially when you consider the fact that Spectra is already running up against crippling processing and data-handling limitations. 

 

The crippling processing computation limitations on the Rafale only exist in your mind. If you'd care to learn a bit about the aircraft you so regularly make fun of, you'd know that the Rafale uses what they call the MDPU (Modular Data Processing Unit), a common processing pool that is shared by all Rafale systems (i.e. radar, ECM...).

 

The F-22 & F-35 uses exactly the same architecture, in this case it's called CIP (Common Integrated Processor). In fact, all these planes now use commercial-derivatives processors of the same family (PowerPC) except for the early blocks F-22 (uses intel i960) and Rafale F1 (no MDPU, I believe the processors used were SPARC based).

 

In the Rafale case, there are up to 18 processing boards in a single MDPU and processing power can be tailored to the aircraft configuration (e.g. the test birds that have been upgraded with the RBE2 AA have two extra MDPU boards).

 

Each board uses a PPC 740 processor running at 200MHz, with a usable processing capacity of ~65 MIPS (MDPU architecture includes a multi-level software stack to hide the hardware caracteristics, making future upgrade easier but eating in the CPU raw power + processes are likely run with a 1+1 redundancy to avoid any interruption in case of a board failure). A Rafale with 18 boards would have a little over 1000 MIPS, or 1.5x the F-22 CPI official processing power. So both systems seem to have similar levels of processing power.

The same architecture (and principle) is used in some of the more recent heavyweight torpedoes, and IIRC, its a variation of the well worn and often abused term of being "distributed"- although they are CPU's per se, they're actually and more accurately RISC chips, so the processing power is inherent more so in the code/algorithms/routines programmed into those RISC chips..- and then coupled to the performance of that chip.  The i960 used to be a favourite processor for onboard RAID controllers about 12 years ago.

The problem for the early Block releases of the F-22 (now that you've pulled it into the open) is that milgrade 960's are hard to come by. Unless I'm way off base, the cooling requirements for thise multi processor SPARC daughter boards is huge - probably 70+% bigger in comparitive real estate.  Depending on your philosophy, then there is an efficiency issue.

I was also under the impression that one of the probs for Rafale and her piggy backed processors was that they generated stacks of heat, and the only way to get around that was via cryo cooling as not enough air could circulate against what was generated. 

This is bordering on vague recollection for me, so I could be a tad off base 

 

Given Intel discontinued the product family that's hardly surprising.

 

One story I've heard about the i960 used in the F-22 is that Intel got the market by promising to deliver a chip that was much larger than the competition and cost a fortune to manufacture and required a dedicated line (at the time the F-22 program was going strong and the Comanche was supposed to use the same chips too so it must have looked like a very lucrative market).

 

But unfortunately by the time the F-22 went online, the i960 and its specific manufacturing process were long outdated, and  Intel did not want to keep precious fab space occupied by what had become a niche product. 

 

I was also under the impression that one of the probs for Rafale and her piggy backed processors was that they generated stacks of heat

which does invite the rhetorical question of how many concurrent tasks per second the system can actually do if its not cuasing the processing block to act like Yellowstone.....   eg Rafale doesn't have the same sensor footprint and onboard integration of (eg the JSF) - so that would seem to make it more likely to have to rely on mounted peripherals with their own onbaord processing.  again, my understanding is that Rafale does have process blocks (comparitively).

hence my laborious and ongoing dialogue to explain why an export F-22 is decidedly unattractive....

 

As far as I can tell the latest blocks F-22 don't use i960 processors so it shouldn't be an issue.

 

Rafale doesn't have the same sensor footprint and onboard integration of (eg the JSF) - so that would seem to make it more likely to have to rely on mounted peripherals with their own onbaord processing

alluding to the fact that Rafale has a lower tolerance for complex activity (comparitively to F-22 or JSF) due to the nature of its design.  eg processing speed, concurrency issues etc....   good for what the french "saw" a few years back, but an integration nightmare as their battlespace management becomes more complex.  they had a few systems problems earlier this year which seemed to indicate that they haven't quite got the comms integration right with other battlespace managers etc....

From your last sentence, does this mean difficulties in sensor fusion especially with external data networks?

 

Fascinating discussion. 

 

If the MDPU is a shared processor resource does this mean that this is a true multiprocessor where processes/programs are not necessarily tied to a specific CPU?