Hacker News Reader

How is Geekbench any good at comparing RISC-V to ARM? Geekbench isn't a native RISC-V application, let alone has the wherewithal to correctly report any basic information like frequency or core count. You haven't even prefaced these either, and drew conclusions from them.

Also, actually searching the chip in question is impossible.

View on HN

There are preview builds of Geekbench that are native RISC-V.

Doesn't mean that it is compiled with Vector support or optimizations, which is going to artificially make the results worse.

Neutral party here, well not so neutral, I up voted you, it's a reasonable question at first blush: however here, you're doubling down with another comment that very clearly indicates you didn't read the link provided

I highly recommend it, most incisive RISC-V article I've read.

Your reply does not address the parent.

Geekbench does indeed not support the Vector extension, and thus yields very poor results on RISC-V.

You reply does not address the great great great grandparent.

RISC-V does indeed not have pipeline reordering.

The Geekbench score thing is a strawman invented to distract from that, no one has mentioned Geekbench except the people arguing it doesn't matter. Everyone agrees, it doesn't matter. So why pound the table about it?

Reply on news.ycombinator.com