Raptor Launching Talos II Lite POWER9 Computer System At A Lower Cost

[grok]

Geforce GT710 is a PCIe 8x card which should get you going, it does three monitors. It has low specifications, uses 19W.

[coder]

Does anyone know whether Linux on POWER uses little endian or big endian?

PowerPC is the only reason I was ever tempted to get a Mac, which became a very plausible option after the switch to the BSD-derived OS X. In the end, I found I couldn't stomach the inflated hardware prices and lack of upgradability.

[madscientist159]

Does anyone know whether Linux on POWER uses little endian or big endian?

PowerPC is the only reason I was ever tempted to get a Mac, which became a very plausible option after the switch to the BSD-derived OS X. In the end, I found I couldn't stomach the inflated hardware prices and lack of upgradability.

It's configurable at runtime. You can literally have LE VMs running on a BE host, or vice versa.

[mongo]

Does anyone have benchmarks on quad precision floats compared to say quadmath?

I have a use case that could benefit from increased precision on intermediate calculations but haven't found anything.

[coder]

It's configurable at runtime. You can literally have LE VMs running on a BE host, or vice versa.

So, I guess what does Ubuntu for Power use (since that's what I'd probably run)?

Do any popular Linux distros (i.e. for Power) differ, in this regard?

[chithanh]

It's the SAS controller. Sadly the associated PCIe controller is locked into a fixed configuration of x8 and x8 in silicon.

Bummer. Thanks for the answering my questions though.

So, I guess what does Ubuntu for Power use (since that's what I'd probably run)?

Do any popular Linux distros (i.e. for Power) differ, in this regard?

Linux distros traditionally have ppc64 big endian ports.
More recently, ppc64le support has been added to a couple of distros and is mostly working. I think Debian/Ubuntu already have ppc64le images for download.

[coder]

Linux distros traditionally have ppc64 big endian ports.
More recently, ppc64le support has been added to a couple of distros and is mostly working. I think Debian/Ubuntu already have ppc64le images for download.

Thanks. I have more confidence in little endian, since that's what x86 uses, as well as Android on ARM.

TBH, I don't even know what happens when you cast a native word-sized pointer to pointer to a smaller type, on big endian machines. My intuition says it's going to be pointing at the most-significant part of the word, but maybe the compiler will offset it? I can just imagine things like that breaking a lot of code - especially as fewer and fewer programmers have to deal with big endian, these days.

[chithanh]

Presumably the same would happen as when you cast an int64 to an int32: the most significant bits will be cut off.

[coder]

Presumably the same would happen as when you cast an int64 to an int32: the most significant bits will be cut off.

I don't have any truly authoritative links to post, but my limited web research clearly suggests that the effect of pointer casting will change, based on endianness.

[chithanh]

my limited web research clearly suggests that the effect of pointer casting will change, based on endianness.

Ok, so you say that the following C code:

#include <stdlib.h>
#include <stdint.h>
#include <stdio.h>

int main() {
        void *foo = malloc(1);
        uint32_t bar = (uint32_t) foo;
        uint32_t baz = (uint32_t)(uint64_t) foo;
        printf("%s\n", bar == baz ? "true" : "false");
}

Will produce different output depending on the endianness of the host system? I don't have a BE system to check right now, will maybe try later.