Jump to content

This CPU is FREE!

TannerMcCoolman

@TannerMcCoolman At 10:14 [0] you show a picture which claims the C920 "supports Vector1.0" and is therefor "compatible with RVA23 profile".
There is no source given for this. After much searching which turned up empty, I accidentally found the site from which the screenshot was apparently taken. [1]

 

This information is wrong but in an interesting way.

 

The Milk-V Pioneer product page [2] clearly states the SG2042 being a "64-core C920, RVV 0.71" [sic].
This is further confirmed in a recent paper published in ACM [3][4]:
"Version 0.7.1 of the vectorisation standard extension (RVV v0.7.1) is supported" (Section 2.1).

 

Interestingly, the C920 is not listed on T-Heads website, [5] neither on the english nor chinese version.
However, this authentic looking spec sheet [6] turned up in my search which also confirms RVV 0.7.1.

 

Recently there's been news articles [7][8] which claims C920 to support RVV 1.0.
So it seems to me, T-Head is trying to quietly relaunch the C920 with different specs.

 

 

[0] https://youtu.be/vaMxTSm53UU?t=614
[1] https://www.xrvm.com/product/xuantie/4237560868973383680
[2] https://milkv.io/pioneer
[3] https://doi.org/10.1145/3624062.3624234
[4] https://arxiv.org/abs/2309.00381
[5] https://www.t-head.cn/
[6] https://occ-intl-prod.oss-ap-southeast-1.aliyuncs.com/userFiles/13559201667542738675/T-Head%20XuanTie%20C920%20Processor%20Datasheet.pdf
[7] https://www.reddit.com/r/RISCV/comments/1817v41/xuantie_c920_rvv_10_c907_r910_released_the_first/
[8] https://www.reddit.com/r/RISCV/comments/1601xv9/c920_with_rvv_10_announced/

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×