4 ms·

I've not heard of it, presumably because it didn't go anywhere 15 years ago. Is there any more background?

by gnufx 6y ago

aakshintala 6y ago
Keep in mind this was an academic project. The project ran its course, produced many dissertations, and more importantly successful researchers and engineers, and inspired many others to look at FPGAs and dataflow. It wasn't ever meant to be about a product or building a company... I would counter that the project was immensely successful with the stated academic goals: as proof, the Microsoft Azure AI HW division (formerly people responsible for project catapult) was almost entirely bootstrapped by many of the same people who worked on the TRIPS project. Here's Doug Burger providing a small retrospective on the project: https://www.microsoft.com/en-us/research/podcast/clouds-catapults-life-end-moores-law-dr-doug-burger/?ocid=msr_podcast_dburger_profile https://www.microsoft.com/en-us/research/podcast/clouds-cata...
- borramakot 6y ago
  You say formerly catapult- is catapult no longer being developed?
  - aakshintala 6y ago
    I don't work for Microsoft or have any real knowledge here, but my understanding (from the grapevine) is that the team essentially was given a much larger purview. So in a sense Catapult is evolving?
- gnufx 6y ago
  I didn't necessarily mean building a product, or to put it down; I just wasn't aware of influences, and I'm close to a home of dataflow. Thanks for the background.
  - aakshintala 6y ago
    My apologies if I came off unnecessarily reactive.
_chris_ 6y ago
Microsoft and I believe Qualcomm research took this and did two follow-on versions. They gave a joint keynote talk at ISCA (2018?) on their experiences, but I can't seem to find any slides or video on it. I recall them saying it looked to give them roughly the same performance as state-of-the-art processors with ~30% less power; but that's a hard gamble for the $100Ms required to get to a product (especially if it requires a new ISA).
- gnufx 6y ago
  Thanks for the info.

I've not heard of it, presumably because it didn't go anywhere 15 years ago. Is there any more background?

by gnufx 6y ago

I've not heard of it, presumably because it didn't go anywhere 15 years ago. Is there any more background?

Keep in mind this was an academic project. The project ran its course, produced many dissertations, and more importantly successful researchers and engineers, and inspired many others to look at FPGAs and dataflow. It wasn't ever meant to be about a product or building a company... I would counter that the project was immensely successful with the stated academic goals: as proof, the Microsoft Azure AI HW division (formerly people responsible for project catapult) was almost entirely bootstrapped by many of the same people who worked on the TRIPS project. Here's Doug Burger providing a small retrospective on the project: https://www.microsoft.com/en-us/research/podcast/clouds-catapults-life-end-moores-law-dr-doug-burger/?ocid=msr_podcast_dburger_profile https://www.microsoft.com/en-us/research/podcast/clouds-cata...

borramakot 6y ago
You say formerly catapult- is catapult no longer being developed?
- aakshintala 6y ago
  I don't work for Microsoft or have any real knowledge here, but my understanding (from the grapevine) is that the team essentially was given a much larger purview. So in a sense Catapult is evolving?
gnufx 6y ago
I didn't necessarily mean building a product, or to put it down; I just wasn't aware of influences, and I'm close to a home of dataflow. Thanks for the background.
- aakshintala 6y ago
  My apologies if I came off unnecessarily reactive.

_chris_ 6y ago

Microsoft and I believe Qualcomm research took this and did two follow-on versions. They gave a joint keynote talk at ISCA (2018?) on their experiences, but I can't seem to find any slides or video on it. I recall them saying it looked to give them roughly the same performance as state-of-the-art processors with ~30% less power; but that's a hard gamble for the $100Ms required to get to a product (especially if it requires a new ISA).

gnufx 6y ago
Thanks for the info.