Watch AMD's Computex 2021 Keynote With PCWorld
video description
Date: 2022-03-15
Related videos
Comments and reviews: 7
Eternalduoae
Actually, what i think is the correct interpretation of FSR support - although they showed it running on a GTX 1060. AMD have categorically stated that support for the feature -MUST- come from the card vendor.
So, I'm pretty sure that Alaina's and Adam's interpretation is not correct. FSR won't -just work- on Geforce 10 through 30 series. It needs driver support.
What i believe they're implying is that the architecture/ instructions available on these generations of cards support the compute operations required to perform the upscaling - hence the Polaris and Vega support.
For me, AMD did this comparison as an FU to Nvidia by making their fans (and owners) believe that the feature works and is supported on GTX and RTX cards. which is a pretty scummy move, IMO.
From the disclaimers in the presentation slides and on AMD's website:
-AMD does not provide technical or warranty support for AMD FidelityFX Super Resolution enablement on other vendor's graphics cards. GD-187-
reply
Actually, what i think is the correct interpretation of FSR support - although they showed it running on a GTX 1060. AMD have categorically stated that support for the feature -MUST- come from the card vendor.
So, I'm pretty sure that Alaina's and Adam's interpretation is not correct. FSR won't -just work- on Geforce 10 through 30 series. It needs driver support.
What i believe they're implying is that the architecture/ instructions available on these generations of cards support the compute operations required to perform the upscaling - hence the Polaris and Vega support.
For me, AMD did this comparison as an FU to Nvidia by making their fans (and owners) believe that the feature works and is supported on GTX and RTX cards. which is a pretty scummy move, IMO.
From the disclaimers in the presentation slides and on AMD's website:
-AMD does not provide technical or warranty support for AMD FidelityFX Super Resolution enablement on other vendor's graphics cards. GD-187-
reply
Peter
Here a little celebration of stacked cache. Most of the slowdown of CPU-cores is being caused by wasted cycles (NOP's) because data is not fetched on time. 'Moving' data (basically high or low voltages on a trace) is a slow process, especially when that data comes from the RAM. This stacked cache is going to improve performance noticeably, especially because AMD seems to have kept the latency under control by having many direct copper to copper connections. I am quite excited about having one of those CPU's. Intel talks about it, AMD delivers it.
reply
Here a little celebration of stacked cache. Most of the slowdown of CPU-cores is being caused by wasted cycles (NOP's) because data is not fetched on time. 'Moving' data (basically high or low voltages on a trace) is a slow process, especially when that data comes from the RAM. This stacked cache is going to improve performance noticeably, especially because AMD seems to have kept the latency under control by having many direct copper to copper connections. I am quite excited about having one of those CPU's. Intel talks about it, AMD delivers it.
reply
Tallyn's
I really enjoy you're analysis when these events come up, thank you for you're hard work. A lot of YouTubers have an obvious bias slant towards one company or another and I appreciate your fair and equal coverage with out any flag planting.
reply
I really enjoy you're analysis when these events come up, thank you for you're hard work. A lot of YouTubers have an obvious bias slant towards one company or another and I appreciate your fair and equal coverage with out any flag planting.
reply
Calvin
Did you really not consider people with Tesla's having their kids play the games while they still drive? There is more than one seat in a car lol can you imagine people hacking Mario Party into the center console for those family trips!
reply
Did you really not consider people with Tesla's having their kids play the games while they still drive? There is more than one seat in a car lol can you imagine people hacking Mario Party into the center console for those family trips!
reply
Erik
Dr. Su understands the actual engineering behind their products and that allows her to present their products on a level that everyone can understand.
AMD was incredibly smart to appoint a former engineer as their CEO.
reply
Dr. Su understands the actual engineering behind their products and that allows her to present their products on a level that everyone can understand.
AMD was incredibly smart to appoint a former engineer as their CEO.
reply
Peter
The Samsung-news is rather old, at least 3 years old. That is, it was well known that AMD worked together with Samsung for this purpose, now they just step it up and they make it official.
reply
The Samsung-news is rather old, at least 3 years old. That is, it was well known that AMD worked together with Samsung for this purpose, now they just step it up and they make it official.
reply
Eternalduoae
So, i was right when i asked Scott if it made sense to scale RDNA 2 down the stack. Navi 24 just doesn't have infinity cache at all.
reply
So, i was right when i asked Scott if it made sense to scale RDNA 2 down the stack. Navi 24 just doesn't have infinity cache at all.
reply
Add a review, comment
Other channel videos
