True, I still find it a bit sad though, if that’s the final case.
I value function over form too, but it looks like parts of the front panel are made from cheap plastic. Especially the top panel and its needlessly chrome-coloured plastic power button.
It looks like Form over Function to me, and not a Form which is particularly appealing.
Well if its a mock up the whole thing is probably the cheapest thing they could get for Demo.
My dad was a 35 year IBMer and I used to get excited about RISC architecture.
But after multiple nothingburgers over the years, most notably PReP architecture and rumors of PowerPC 604 based motherboards made by Tyan back in the mid 90s, I got tired of getting let down. Believe me, the thought of OS/2 on a 604 PowerPC at the time was so intoxicating that I was practically delirious.
As much as I’d like to see POWER be a viable competitor, I just don’t think that it ever will be in today’s era of cheap and good enough x86 compatible systems. The only RISC system I see with any potential to unseat x86 is ARM, and it’s going to have to win a race to the bottom… ultra cheap computing (Raspberry Pi prices).
But there is nothing wrong with a little day dreaming. I’ve already read all about this and watched too many videos on it. I just can’t get excited or be bothered to spend any money on products that probably will never be available in any meaningful or cost effective quantity. Let me know when you got your TALOS system up an running. Maybe you can reignite the fire in me.
Until then… good luck.
I hope so.
I also recently came across someone remarking that a post on twitter from February, describes a motherboard near identical to Talos II’s as an “IBMPowerSystems developers board”.
This makes me wonder if Raptor Computing/Engineering is actually making/contracting the board themselves, or if are they just reselling a developer board. This would also explain the chopped down 4-core Sforza’s they are initially shipping with it.
As I was writing this I noticed that less than an hour ago they announced a press release, and started using the “RaptorCompSys” twitter account, which had been sitting dormant until now. As of now I still see no mention of the board being a developer board, though the 4U server is labelled as a development platform the workstation is not. The plot thickens…
Weird sidenote from looking into this stuff, apparently there is a Twitter account that tracks when trademarks are filed. So apparently, Raptor Computing Systems was also filed back in September 2016, when the Twitter account was made.
you can bypass the security coprocessor with a kernel patch that points to a modified rom of your choice.
Search wolf0 on literally any mining forum. He’s been talking about it for more than a week.
Well lets keep off of curreeencies. This is a thread about a certain computer so lets keep it there before it derails.
Sorry about that. Moving Vega Firmware Signing info here:
So I feel like for some people on this forum they may not know what the differences are… So lets talk about little endian vs big endian shall we?
Little Endian and Big endian are two ways of storing multi-byte data-types ( int, float, etc) in computers. In a sense this is how the computer stores data in its processor registers, which can hold a variety of data types from instructions to storage addresses to stack pointers.
In Little Endian machines, last byte or least significant bit of binary representation of the multi-byte data-type is stored first while in Big Endian machines, first byte or most significant bit of binary representation of the multi-byte data-type is stored first.
Suppose an integer is stored as 4 bytes(32-bits), then a variable y with value 0xdeadbeef( Hexa-decimal representation) is stored as four bytes 0xDE, 0xAD, 0xBE, 0xEF, on Big-endian while on Little-Endian (Intel x86), it will be stored in reverse order as follows 0xEF 0xBE 0xAD 0xDE.
Example: How 0x1234567 is stored at memory location 0x100-ox103 on any standard machine at home?
Here is a piece of code from my ECE330 class that I wrote to do exactly that for a test given zero help to test endianness the third day of class. Execute it at home on a linux machine prefereably with all the build tools LOL… Or windows whichever you may prefer the operating system is completely irrelevant
#include <stdio.h>
/* function to show bytes in memory, from location start to start+n*/
void show_mem_rep(char *start, int n){
int i;
for (i = 0; i < n; i++)
printf(" %.2x", start[i]);
printf("\n");
}
/*Main function to call above function for 0xba5eba11*/
int main(){
int i = 0xba5eba11;
show_mem_rep((char *)&i, sizeof(i));
return 0;
}
How to remember?
Well think about this. If you have two fishes. In big endian you have the whole fish but lets try to think about it as two halves of the fish in the correct order. representing each bit. Now in little endian you cut the fish in half and put the tail first then start counting. See the difference? Draw it for yourself all you gotta remember is that in Little Endian, things are stored in reverse order.
In “Little Endian” form, assembly language instructions for picking up a 1, 2, 4, or longer byte number proceed in exactly the same way for all data storage formats. The process goes like this… first pick up the lowest order byte at offset 0. Also, because of the 1:1 relationship between address offset and byte number (offset 0 is byte 0), multiple precision math routines are correspondingly easier to write. Armed with that knowledge we can see why we write most mathematical routines in little endian.
In “Big Endian” form, by having the high-order byte come first, you can always test whether the number is positive or negative by looking at the byte at offset zero which is incredibly useful because you don’t have to know how long the number is, nor do you have to skip over any bytes to find the byte containing the sign information of the data or numerical instruction set. The numbers are also stored in the order in which they are printed out, so binary to decimal routines are particularly efficient and easier to read by humans
Glad you asked…
Adobe Photoshop – Big Endian
BMP (Windows and OS/2 Bitmaps) – Little Endian
DXF (AutoCad) – Variable
GIF – Little Endian
IMG (GEM Raster) – Big Endian
JPEG – Big Endian
FLI (Autodesk Animator) – Little Endian
MacPaint – Big Endian
PCX (PC Paintbrush) – Little Endian
PostScript – Not Applicable (text!)
POV (Persistence of Vision ray-tracer) – Not Applicable (text!)
QTM (Quicktime Movies) – Little Endian (on a Mac!)
Microsoft RIFF (.WAV & .AVI) – Both
Microsoft RTF (Rich Text Format) – Little Endian
SGI (Silicon Graphics) – Big Endian
Sun Raster – Big Endian
TGA (Targa) – Little Endian
TIFF – Both, Endian identifier encoded into file
WPG (WordPerfect Graphics Metafile) – Big Endian (on a PC!)
XWD (X Window Dump) – Both, Endian identifier encoded into file
And so on
If you ever find yourself in need of reversing the order of a multibyte integer to suit your liking… there are many methods to doing so
It is pretty easy to reverse a multibyte integer if you find you need the other format. A single function can be used to switch from one to the other, in either direction. A simple and not very efficient version might look as follows:
Function Reverse (N:LongInt) : LongInt ;
Var B0, B1, B2, B3 : Byte ;
Begin
B0 := N Mod 256 ;
N := N Div 256 ;
B1 := N Mod 256 ;
N := N Div 256 ;
B2 := N Mod 256 ;
N := N Div 256 ;
B3 := N Mod 256 ;
Reverse := (((B0 * 256 + B1) * 256 + B2) * 256 + B3) ;
End ;
A more efficient version that depends on the presence of hexadecimal numbers, bit masking operators AND, OR, and NOT, and shift operators SHL and SHR might look as follows:
Function Reverse (N:LongInt) : LongInt ;
Var B0, B1, B2, B3 : Byte ;
Begin
B0 := (N AND $000000FF) SHR 0 ;
B1 := (N AND $0000FF00) SHR 8 ;
B2 := (N AND $00FF0000) SHR 16 ;
B3 := (N AND $FF000000) SHR 24 ;
Reverse := (B0 SHL 24) OR (B1 SHL 16) OR (B2 SHL 8) OR (B3 SHL 0) ;
End ;
There are certainly more efficient methods, some of which are quite machine and platform dependent as @FaunCB pointed out the Power Architecture has its own way of doing so that is optimized for its platform… Use what works best for your situation
Source:
The case, if you didn’t recognize it, is a Supermicro 4U/tower case, same or similar to this one. I’ve worked with one of these before and although it might not have a “premium” appearance or feel, I don’t have any particular complaints about the build quality. Overall for a white box server/workstation chassis it’s a pretty solid choice.
It wouldn’t make much sense at this stage for Raptor to be investing resources into a custom case design, especially after they failed to secure crowdfunding for their previous attempt at a POWER8 system.
That is good to hear; also, now I can see read/see about the case, thank you!
Edit: looks most like SC747.
I find it interesting the design decision to go with a Power PC derived platform over developing the same compute as an add in card. After all you need application support for a non-86 architecture anyway and the PCIe bus is faster than basically any networking solution, so placing their POWER cores onto a card that could be used in much the same way as compute\commercial graphics cards already are would seem a more widely deploy-able setup than dedicated boxes, so I’d love to hear why they chose to go the other way.
The point is to be able to go from power on to OS without needing to use any proprietary code at all.
OCC code (on chip controller - always running)
CPU code
SBE -> hostboot -> skiboot -> petitboot -> OS
all of that is open source, and if you use OpenBMC, then you can remotely power on and off and still never need to trust a black box.
Good point, my head was in a increased efficiency for some task vs 86x mindset not a open standards mindset.
The company behind it has a one of their primary goals of producing an open system so it makes sense in that context.
I’m surprised there isnt more talk about this here considering the number of people that are always complaining about big companies doing bad things with closed propitiatory systems.
Also, you would be needing to slot RAM into a PCIe device, or else have RAM soldered to it like a GPU. Not to mention, you would also need a host system CPU (which would increase cost) and have to deal with carting data between card-RAM and system-RAM.
In short, you would be cajoling a CPU to work like a GPU, and that seems like a lot more development than just making an EATX mainboard. It might be a cool idea, but I don’t see what purpose it would serve.
Other than a very select few no one know what this thing is. With that comes the “what to do we do with such a thing” problem. To most people it is as if you dropped a UFO’s navigation computer on there desk and said “make that run minesweeper”.
I am one of these people. It certainly looks interesting but it is very expensive, not aimed at average users, and seemingly has absolutely no community to fall back on (though I have no idea) unlike all current systems.
Because that would make too much sense. A POWER coprocessor card ala a Xeon Phi, or Tesla would be a lot cheaper to produce, and would be a better way to introduce new buyers to the technology.
The desire to totally unseat x86 as a platform is clouding their vision. It’s absurd.
Nice troll.