One possibility is that it could have a memory controller designed to work with either GDDR5 or GDDR5X. Parts that are low-end, or ultra-low-power would be designed to use standard GDDR5 at lower core clocks, while the 75W desktop card would use GDDR5X and higher core and memory clocks to allow the chip to stretch its legs.
We've seen dual memory controllers in the past, especially on small chips; Cape Verde works with both DDR3 and GDDR5, for instance. And GDDR5X is supposed to be a relatively simple addition to the GDDR5 standard, this shouldn't cost much extra die space.
Even 1st generation GDDR5X is supposed to have 12 Gbps data rates (eventual target is 16 Gbps, and Micron claims to have hit 13 already). A 128-bit bus with 12 Gbps GDDR5X would provide 192 GB/sec of memory bandwidth, which is more than the current Tonga cards have on tap.
Micron says mass production of GDDR5X starts this summer. If Apple needs Polaris 10 sooner for the MacBook Pro refresh, standard GDDR5 would still let the chip offer unparalleled performance per watt, and slower core clock rates would make a memory bandwidth bottleneck less of an issue.
You forget some new posters claimed that Fury X will be 40% faster then Titan X and they diapered after Fury X release. Lack of hype control from AMD.
http://semiaccurate.com/forums/showpost.php?p=255514&postcount=428According to this:
http://www.pcr-online.biz/news/read...july-with-amd-polaris-gpus-expected-q3/037923
Pascal is only a couple months from launch, and Polaris will be Q3...
I seriously doubt this as Pascal GPUs only *just* started showing up on Zauba, which would suggest they are at least 6 months from launch, whereas Polaris should be able to launch in 3 or 4 months, going by both companies' historic timelines (from first Zauba listing to launch).
This is not the only claim I've seen that Pascal will beat Polaris to market, but the traditional leak sources strongly suggest otherwise...
Thoughts?
There are options between 128 bit and 256 bit for the memory bus. A bit better memory compression with a 192 bit bus allowing 3GB & 6 GB for the slowest new cards.
According to this:
http://www.pcr-online.biz/news/read...july-with-amd-polaris-gpus-expected-q3/037923
Pascal is only a couple months from launch, and Polaris will be Q3...
I seriously doubt this as Pascal GPUs only *just* started showing up on Zauba, which would suggest they are at least 6 months from launch, whereas Polaris should be able to launch in 3 or 4 months, going by both companies' historic timelines (from first Zauba listing to launch).
This is not the only claim I've seen that Pascal will beat Polaris to market, but the traditional leak sources strongly suggest otherwise...
Thoughts?
GDDR5X is not going to be used on low end GPUs where cost is a major factor. The production volumes of GDDR5X will be low this year as its a brand new technology which will start to ramp in Q3 2016. I expect to see GDDR5X in high end GPUs like GP104 where the products are sold at high prices and the increase in production costs can be easily absorbed. AMD also mentioned that Polaris supports GDDR5 and HBM. There was no mention of GDDR5X.
http://wccftech.com/amd-radeon-technologies-group-plans-to-spice-up-gdc/
AMD lately likes to play with public through Twitter.
You've got to look at the total bill of materials. If GDDR5X allows halving the bus width compared to GDDR5, it might be worth it even if the chips cost 50% more (I'm just throwing that number out there randomly, I have no idea what the real figures are). A wider memory bus means more power consumption, more PCB traces, more die space, and more physical chips. All that costs money. A few extra bucks on the RAM ICs themselves could easily be offset by this.
This is a bad comparison. GDDR5X is an extension of GDDR5, not a totally new memory technology and thus should be applied much faster than traditional transitions indicate.The key here is new memory technologies take time to ramp production volume and get yields to a point where it make sense at lower price points. GDDR5 was launched in June 2008 with HD 4870. After 15 months we saw the next product with GDDR5 (HD 5870). I expect something similar with GP104 launching in Q4 2016. Maybe we will see GDDR5X at lower price points (USD 200 - USD 300) in late 2017 or early 2018.
This is a bad comparison. GDDR5X is an extension of GDDR5, not a totally new memory technology and thus should be applied much faster than traditional transitions indicate.
The key here is new memory technologies take time to ramp production volume and get yields to a point where it make sense at lower price points. GDDR5 was launched in June 2008 with HD 4870. After 15 months we saw the next product with GDDR5 (HD 5870). I expect something similar with GP104 launching in Q4 2016. Maybe we will see GDDR5X at lower price points (USD 200 - USD 300) in late 2017 or early 2018.
Is that the only two options you can see. 1 1/2 - 2 yrs as you proposed or all at once.GDDR5X might be an extension of GDDR5. But that does not mean GDDR5X will be able to replace all GDDR5 cards from the price points of USD 100 - 1000. No new technology ramps like that. You will see GDDR5X launch at the high end and then gradually move down the stack in a year or so time. By your logic we should see all Pascal and Polaris cards use GDDR5X. But thats not going to happen.
http://wccftech.com/amd-radeon-technologies-group-plans-to-spice-up-gdc/
AMD lately likes to play with public through Twitter.
2GB. Ugh. Those are probably very entry level cards.
Shitai, on that screenshot I do not see anywhere GDDR5 mention on Baffin XT. So why count it as GDDR5 chip?
Weston X3
Banks S3.
Baffin G5.
Where is GDDR5 coming from?
Shitai, on that screenshot I do not see anywhere GDDR5 mention on Baffin XT. So why count it as GDDR5 chip?
Weston X3
Banks S3.
Baffin G5.
Where is GDDR5 coming from?
Its from the article. And you should correct your typo.
The speculation was G5 is a shortening of GDDR5.
Then what's X3 and S3?