- Jun 6, 2013
- 1,164
- 406
- 136
With AMD becoming even more irrelevant with every product paperlaunch plus all the previous flops that they already accumulated in the last half decade, is probable that in less that two years we will see two possible outcomes: AMD could go bankrupt and dissapear along with their x86 license, or maybe there could be a meteor shower of miracles and Zen becomes a success that allows them to somehow recover and become competitive. But these days I'm not that optimist, and would prepare for the worst.
The main problem is that we already are getting an early taste of what would happen if Intel becomes an absolute monopoly with no competition (Through what AMD provides currently can't be called that...). They don't bother to compete with themselves, just milking the cash cow with market segmentation. Intel decided for the Skylake generation to not allow using Xeons E3 on Desktop Motherboards as Core i5/i7 replacements (Which could have put a cap on the Core i7 6700K price gouging), and we also saw in the last days that Intel has the last word over what features Motherboard makers may enable, as it seems that Intel was able to force them to kill Base Clock overclock on non-K Processor and/or non-Z Chipsets.
So at this point, I'm rather interesing in how viable is actually jumping out of x86 before it becomes too late and we all do Intel bidding...
The problem is that you used to only have architecture variety on big Servers and mainframes (Anyone remembers the DEC Alpha?), which obviously no one has the budget for. The alternative closer to mainstream and reasonable budgets is ARM, but these aren't your traditional do it yourself Desktop system and you usually see it in devices like Tablets.
When the AMD Opteron A1100 was announced, I was very interesed in it since what it could achieve was providing a Desktop PC-like approach to a non-x86 architecture, so I could do something like, purchase a standard ATX Case and Power Supply, DDR3 RAM, a SATA SSD, and a Motherboard with the AMD Seattle SoC, then I would be able to assemble an ARM system whose look and feel is much closer to a traditional Desktop, if not identical, than an ARM Tablet or other consumer gadget using their CPU ISA.
So far, this idea didn't seem to go anywhere since as far that I know, there are still no Opteron A1100 benchmarks, nor you can purchase it.
Today we have a new contender. Phoronix announced that there is a small business interesed in building systems around OpenPOWER. The guys that are in that project have experience with some ports of Coreboot, including to at least one AMD Server Motherboard. I don't know who commisioned the port, but they are the only small business that I'm aware of that does jobs like those.
Obviously that the main advantage is that you can step out of x86 with it, so you don't have to bow to Intel. You also have an Open Source Firmware. I think that even the Hardware side is Open Source, as IBM with the OpenPOWER iniciative is actually providing for the POWER8 Processor schematics, but I'm not entirely sure if an end user can get them (Sun with OpenSPARC did so if I recall correctly).
As we know, BIOSes are full of obscurity, and the maintenance via updates is done by the Motherboard maker for as long as they want to do so, and to fix or break whatever they care about. Product is EOL? Sorry, you're screwed (Unless you happen to have one of the very few Motherboards supported by Coreboot). There are too many things that we don't have control of in the x86 platform, and we're losing ground every generation. So, enter OpenPOWER, a possible exit to the x86 platform assuming that ARM isn't viable enough to do it.
There is an issue: The announced 3100 U$D price tag, and that's assuming there are enough interesed people to make a production run of the prototype system. Freedom is not for people in a budget, so I don't think that you will see anyone replacing his everyday system with one of these (And forget about Windows, or nearly all games for that matter...). But is extremely interesing to see that at least x86 alternatives that could eventually become mainstream, assuming that price somehow becomes acceptable.
The main problem is that we already are getting an early taste of what would happen if Intel becomes an absolute monopoly with no competition (Through what AMD provides currently can't be called that...). They don't bother to compete with themselves, just milking the cash cow with market segmentation. Intel decided for the Skylake generation to not allow using Xeons E3 on Desktop Motherboards as Core i5/i7 replacements (Which could have put a cap on the Core i7 6700K price gouging), and we also saw in the last days that Intel has the last word over what features Motherboard makers may enable, as it seems that Intel was able to force them to kill Base Clock overclock on non-K Processor and/or non-Z Chipsets.
So at this point, I'm rather interesing in how viable is actually jumping out of x86 before it becomes too late and we all do Intel bidding...
The problem is that you used to only have architecture variety on big Servers and mainframes (Anyone remembers the DEC Alpha?), which obviously no one has the budget for. The alternative closer to mainstream and reasonable budgets is ARM, but these aren't your traditional do it yourself Desktop system and you usually see it in devices like Tablets.
When the AMD Opteron A1100 was announced, I was very interesed in it since what it could achieve was providing a Desktop PC-like approach to a non-x86 architecture, so I could do something like, purchase a standard ATX Case and Power Supply, DDR3 RAM, a SATA SSD, and a Motherboard with the AMD Seattle SoC, then I would be able to assemble an ARM system whose look and feel is much closer to a traditional Desktop, if not identical, than an ARM Tablet or other consumer gadget using their CPU ISA.
So far, this idea didn't seem to go anywhere since as far that I know, there are still no Opteron A1100 benchmarks, nor you can purchase it.
Today we have a new contender. Phoronix announced that there is a small business interesed in building systems around OpenPOWER. The guys that are in that project have experience with some ports of Coreboot, including to at least one AMD Server Motherboard. I don't know who commisioned the port, but they are the only small business that I'm aware of that does jobs like those.
Obviously that the main advantage is that you can step out of x86 with it, so you don't have to bow to Intel. You also have an Open Source Firmware. I think that even the Hardware side is Open Source, as IBM with the OpenPOWER iniciative is actually providing for the POWER8 Processor schematics, but I'm not entirely sure if an end user can get them (Sun with OpenSPARC did so if I recall correctly).
As we know, BIOSes are full of obscurity, and the maintenance via updates is done by the Motherboard maker for as long as they want to do so, and to fix or break whatever they care about. Product is EOL? Sorry, you're screwed (Unless you happen to have one of the very few Motherboards supported by Coreboot). There are too many things that we don't have control of in the x86 platform, and we're losing ground every generation. So, enter OpenPOWER, a possible exit to the x86 platform assuming that ARM isn't viable enough to do it.
There is an issue: The announced 3100 U$D price tag, and that's assuming there are enough interesed people to make a production run of the prototype system. Freedom is not for people in a budget, so I don't think that you will see anyone replacing his everyday system with one of these (And forget about Windows, or nearly all games for that matter...). But is extremely interesing to see that at least x86 alternatives that could eventually become mainstream, assuming that price somehow becomes acceptable.