Parts, Boards and the Rotator


Ok, now I'm confused. Are you SERIOUS serious that you need 100k? Or are you still carrying on the joke? Because that's not the kind of thing to be left as a footnote. If such a major investor is pulling out that's big trouble.

I never said I needed it.
I just said an investor with 100k bails out the end of the year and therefore, new investors can come in with 100k.
I had an investment limit with the Pandora, as I can't afford to pay a too high amount if interest rate as well :)

Oh, and in case if anyone wants to know the reason the investor wants his money back:
He supports OpenSource ideas and has a new project where he wants to help out. Similar as he helped out the Pandora years ago.
So he's moving the money to a different open company with a good idea :)
 
Better : Kickstarter for a PowerPC module and the video should contain the sentence "Equipped with the same technology that powered the [insert popular apple product]."

As much as I'm a sucker for pre-Intel Macs I think mentioning Curiosity, Spirit, and Juno would be more compelling.
 
thanks ed for the clarification of this money thingy!

-snip-
i would love to see an lowrisc soc with fpga addon as an possible pyra cpu upgrade.
 
Were any popular Apple products powered by powerPC chips? IIRC the era of PowerPC (around 1994->2004ish??) coincided with a still fired Jobs. Seemed to me that in their day, 68k Apples were more popular than PowerPC ones in their day (smaller market overall probably though), and in some countries, 6502 Apples were probably the most popular, as a percentage of the market. But ARM Apples, be they pod, pad, watch or phone are probably the most popular of all, at least in terms of absolute numbers.

The coolest things I can think of powered by PowerPC are games consoles - the WiiU, Wii, Gamecube and xbox360 all ran those babies.
 
Were any popular Apple products powered by powerPC chips? IIRC the era of PowerPC (around 1994->2004ish??) coincided with a still fired Jobs. Seemed to me that in their day, 68k Apples were more popular than PowerPC ones in their day (smaller market overall probably though), and in some countries, 6502 Apples were probably the most popular, as a percentage of the market. But ARM Apples, be they pod, pad, watch or phone are probably the most popular of all, at least in terms of absolute numbers.

The coolest things I can think of powered by PowerPC are games consoles - the WiiU, Wii, Gamecube and xbox360 all ran those babies.

The iMac, not sure if it beats 68k macs in units sold, but I'm sure it's up there. Next to the iPod it was a huge component of Apples redemption after Jobs returned ~97.
 
Were any popular Apple products powered by powerPC chips? IIRC the era of PowerPC (around 1994->2004ish??) coincided with a still fired Jobs. Seemed to me that in their day, 68k Apples were more popular than PowerPC ones in their day (smaller market overall probably though), and in some countries, 6502 Apples were probably the most popular, as a percentage of the market. But ARM Apples, be they pod, pad, watch or phone are probably the most popular of all, at least in terms of absolute numbers.
Power Mac G5 Was really popular and it ran until late 2005, OS X was originally released on this. It was launched post Steve Jobs return.
 
We had two in our office for rendering, it employed a liquid cooling system since the G5 was just wildly overclocked and screamed along right at the very edge of its thermal envelope, I think that was a big reason Apple ended up migrating to X86.

Anyways, some major failure occurred overnight and I came into work to see the rug underneath the rig stained with the bright green fluid... at least I think it was bright green. Its been a while. Well the wonderful internal design meant the PSU was under the pump assembly for the fluid so It was a total freaking write off. I pulled the drives and we had to trade it in for something else. I think we ended up getting the x86 ones later.
 
@PowerPC-discussion : I actually thought that some of the old and popular Apple products used PowerPC CPUs ... Well, maybe not.
Anyway, I just choose a random group of people spending way too much money on stuff that isn't worth it.
 
For initiators that generate 2D bursts like the display subsystem and IVA-HD however TILER actually is free afaik.
I don't understand this last bit, but I feel somebody should need to explain me too many things here (or I should have to read up too much)
It's actually not that complicated.... I think.

The first thing you need to know is that DDR3 uses a burst length of 8 transfers. Since each DDR3 controller on the omap (it has two) has a 32-bit data bus, that means that every access results in an aligned 8 * 32-bit = 32-byte transfer. If you really only asked for a few bytes, the rest is wasted. Now, imagine what would happen if you'd rotate the framebuffer by scanning it out vertically: you would only need 4 bytes (one pixel), discard the remaining 28, go to the next line and repeat. Simply awful.

Now we introduce TILER. One of the things it does is rearrange the addressing such that a DDR3 burst covers a 4x2 pixel block instead of 8x1. Now you'd waste only 4 bytes per pixel while writing the buffer (in normal orientation) and 12 bytes while reading it (rotated). Total 16 bytes wasted, already a lot better than 28.

The display controller however isn't limited to simple linear bursts. It can ask for 2-dimensional blocks of data from TILER and uses that to ensure it reads complete DDR3 bursts whenever possible, thus avoiding wasted bandwidth. IVA-HD (the video en-/decoding accelerator) has a similar capability. There's unfortunately no way to make such bursts with the cortex-A15, which is why its access through TILER still incurs a penalty.

(Another thing TILER does is distribute the framebuffer across the two DDR3 controllers in a checkerboard pattern, thereby ensuring typical access patterns are nicely divided among the two controllers.)

As far as I can tell it refers to subsystems within the OMAP chip that we're unlikely to be able to use for some time, so I wouldn't worry about it.
Yes, you'll need to SSH in from another machine for the time being, but surely that is only a minor inconvenience. And who knows, maybe one day the display subsystem might work. :p
 
If it would get us the minimum order amounts of stuff we could use for our Linux system, why not?
I know many companies that sell something they don't really like to finance the things they really want to make, mine included.

The Americans wouldn't have a problem getting the minimum order amount for their 4G chips this way for example.
 
Kickstarter for Win10 compatible module after PGS is over? ;)
At this point I thought you were just having a good laugh...

If it would get us the minimum order amounts of stuff we could use for our Linux system, why not?
I know many companies that sell something they don't really like to finance the things they really want to make, mine included.

The Americans wouldn't have a problem getting the minimum order amount for their 4G chips this way for example.
At this point I think clarification is needed.

An X86 SoC module isn't a bad idea.
Windows is a bad idea.

The two can be exclusive from each other.

I like the idea of a Cherry Trail X86 Pyra module (some day) - even if it has to be throttled down to 20% after 2 minutes of continuous 'full-on' use. It's perf/W at any given temp would still be considerably higher than the OMAP 5.

I'm happy - thrilled even - with running Debian. Trying to support Windows X and it's fractured application compatibility on a small volume device would be a nightmare.

Many companies roll a Windows license into the purchase price. Some of them lock the bootloader, some don't. Of those that don't, the general response on Linux is that they won't stop you from supporting yourself.

A hypothetical X86 Pyra could do something similar - come rolled with a solid Linux installation with an unlocked bootloader - if someone wants to fiddle with Windows on it, they're welcome to do so without any official support.

I don't consider Microsoft or Windows to be inherently evil or get into the conspiracy theories about colusion with secret governmental agencies. Linux is simply more stable, more robust, easier to use, easier to maintain. I switched over many years ago and haven't been dissapointed yet.
 
An X86 SoC module isn't a bad idea.
Windows is a bad idea.
The two can be exclusive from each other.
I like the idea of a Cherry Trail X86 Pyra module (some day) - even if it has to be throttled down to 20% after 2 minutes of continuous 'full-on' use. It's perf/W at any given temp would still be considerably higher than the OMAP 5.
I'm happy - thrilled even - with running Debian. Trying to support Windows X and it's fractured application compatibility on a small volume device would be a nightmare.
Many companies roll a Windows license into the purchase price. Some of them lock the bootloader, some don't. Of those that don't, the general response on Linux is that they won't stop you from supporting yourself.
A hypothetical X86 Pyra could do something similar - come rolled with a solid Linux installation with an unlocked bootloader - if someone wants to fiddle with Windows on it, they're welcome to do so without any official support.
I wrote about a Windows compatible module, shipping it with Windows would probably be too much.
An ARM-board is what ED wants to make and imho for good reasons although I also see the appeal of a x86 Pyra module since my DRM-free x86 Linux games collection is quite huge by now.
I don't consider Microsoft or Windows to be inherently evil or get into the conspiracy theories about colusion with secret governmental agencies. Linux is simply more stable, more robust, easier to use, easier to maintain. I switched over many years ago and haven't been dissapointed yet.
I actually do consider Microsoft inherently evil and probably also being into governmental contracts, therefore I don't use it but I don't support Facebook, Google+, McDonalds, Burger King, Vattenfall, Telekom and a bloody lot of other BS either so don't mind me.
 
I wrote about a Windows compatible module, shipping it with Windows would probably be too much.
An ARM-board is what ED wants to make and imho for good reasons although I also see the appeal of a x86 Pyra module since my DRM-free x86 Linux games collection is quite huge by now.

I actually do consider Microsoft inherently evil and probably also being into governmental contracts, therefore I don't use it but I don't support Facebook, Google+, McDonalds, Burger King, Vattenfall, Telekom and a bloody lot of other BS either so don't mind me.
http://lyrics.wikia.com/wiki/Mojo_Nixon:You_Can't_Kill_Me
7th or so verse for hillarious McD reference.

Song was part of the Redneck Rampage CD audio soundtrack. Funny game - great soundtrack.
 
I'd also like a x86 CPU board for my x86 Linux games without needing an emulator at some point and don't see a problem with allowing those that want to try to get Windows running on it to try.

There have been some claims that MS will allow Windows to be free for OEMs on devices smaller than 8 inches but I don't know if that's true and even if it is I'd rather have Linux.
 
Even if MS did allow OEMs free installs of Win10 on devices with small screens, it wouldn't help ED trying to sell CPU boards; the CPU board doesn't have a screen. It might well have the eMMC/NAND flash on it - the micro SD slot is on the peripheral board, but it probably makes more sense to have the eMMC on the CPU board alongside the CPU and RAM.

He might be able to sell complete units with the discounted Windows install, but that wouldn't help us preorderers. Still might not be a bad move though - it could attract people looking for a new laptop who are used to spending >500EUR on such things, but for the time being I'd rather ED focussed on getting the Pyra out at all, and not worry too much about cooling a hungry x86 chip and licensing Windows from MS.
 
Back
Top