The New CXL Standard

[@jaffarbh]

That was brilliant, and answers the questions in my mind before even asking (about using multiple accelerators). Thank you guys

[@SperlingMediaGroup]

Glad it was helpful!

[@echindaplatypus]

Fantastic video, thanks for all your hard work Ed.

[@aliuzel4211]

Great video. Thank you.

[@samc6368]

Nice short tutorial ! I used to work few cubes from Gary at G*, he always explains simplified at any level. Thanks Gary.

[@SperlingMediaGroup]

Glad it was helpful!

[@Alorand]

What I would like to know is if this standard could help multiple GPUs share a workload in a scalable and low latency way that would improve on what is possible with SLI or Crossfire.

[@surajby8089]

Have some queries: 1. How does CXL achieve low latency? Is it by introducing the device biases and multiplexing three different sub protocols? 2. What do you mean by dis-aggregation? How does CXL solve this problem?

[@sumankumarpatra7115]

1. Low latency: In comparison to the existing cache/mem protocols PCIe link is much faster 2. Dis-aggregation of the workload from CPU to accelerators

[@alexisfrjp]

There is nothing special for low latency in the protocol, it's just the way you access data. For a DMA operation, instead of reading the DDR, you read the cache that has a lower read latency. It's a fake low latency characteristic.

[@siddhiL-sd6ru]

thanks

[@chethanm5355]

we can achieve Gen1 --> Gen3 speed without going to Gen5 in CXL ?

[@sumankumarpatra7115]

Yes...But it will, in application, defeat the purpose of high bandwidth

[@chethanm5355]

one query? CXL support Gen4 device ?

[@sumankumarpatra7115]

Yes. Gen3 and above

[@anupganesh2767]

@@sumankumarpatra7115 : Native Gen3 device or it has to be Gen5 degraded to Gen3 ?

[@vinayt7159]

How to know full concept of CXL

[@vinayt7159]

Please suggest any one

[@SperlingMediaGroup]

Here are some additional resources in Semiconductor Engineering's Knowledge Center semiengineering.com/knowledge_centers/standards-laws/standards/compute-express-link-cxl/

[@vinayt7159]

@@SperlingMediaGroup Thank you so much i