WYP
News Guru
Intel's posting some impressive ray tracing performance numbers.
Read more about Intel's ARC Alchemist ray tracing performance.
Read more about Intel's ARC Alchemist ray tracing performance.
Intel details its Ray Tracing graphics performance - Their A770 beats RTX 3060
- Thread starter WYP
- Start date
NeverBackDown
AMD Enthusiast
Pretty good for a first attempt at a GPU and RT. Really shows how bad AMD are doing
Not all that impressive considering that Intel previously compared the A750 to the 3060. For example, Cyberpunk 2077 was about 10% faster on the A750 than 3060. With raytracing, the A770 is about 10% faster than the 3060. So we went up a tier on the Intel side but relative performance is about the same. Intel's implementation therefore seems weaker than NVIDIA's.
NeverBackDown
AMD Enthusiast
It's that much faster in RT. It wasn't referring to rasterization performance.
Avet
Active member
This is amazing. People seem to forget how long it took AMD for Ryzen to start leapfrogging Intel CPUs. And Intel has much, much, much more money and manpower than AMD did when they started CPU battle. Also, they have Tom Petersen and Raja Koduri who actually designed and developed practically all current GPUs.
KingNosser
Well-known member
I am far less optermistic, Intel have lost a lot of money on this and they are really late, by the time anyone see's this in the west if at all then who is going to want to buy them against the 4000 or 7000 series from nvidia and amd.
In reality i expect a limited run of cards and if they don't sell then it'll get shutdown, oem's are the most likely to buy them as they get great deals, but Intel while they might have a lot of man power and funding are not in the biz of losing money.
For battlemage to even get going fully they will already be behind again, they are not about to commit to the road map for a decade to get where they want to be if it doesn't make them money then forget it.
Intel are not going to be the hero, they were just the hero you hoped for.
let alone the drivers they need some kind of internal communication to get them right cause they are all over the place only need to watch a few YT videos to see that rather than cherry picked benchmarks or did we forget about the external reviews they paid for with the cpu's a few years back ?
In reality i expect a limited run of cards and if they don't sell then it'll get shutdown, oem's are the most likely to buy them as they get great deals, but Intel while they might have a lot of man power and funding are not in the biz of losing money.
For battlemage to even get going fully they will already be behind again, they are not about to commit to the road map for a decade to get where they want to be if it doesn't make them money then forget it.
Intel are not going to be the hero, they were just the hero you hoped for.
let alone the drivers they need some kind of internal communication to get them right cause they are all over the place only need to watch a few YT videos to see that rather than cherry picked benchmarks or did we forget about the external reviews they paid for with the cpu's a few years back ?
Rumour has it that Intel will incentivise OEMs to use Arc GPUs.
Software development is pocket change when compared to hardware development, so it's certainly worth it for Intel to continue working on the drivers, in the hope that by the time Battlemage arrives they know well enough how to make a solid driver.
I don't think Intel will cut its lossses unless it assumes that it will continue to remain behind on the hardware development front, which I think is something that Intel rarely assumes.
Which is precisely why it's likely to continue. It's already well into the development of Battlemage and has spent a lot of money on Alchemist. Intel needs, and I think still can, make some money off this. It already ordered a number of wafers, is producing chips, and so it will want to make the best of what it has. It certainly won't be a big consumer success, but I think that Intel will make its best effort to sell Alchemist.
Software development is pocket change when compared to hardware development, so it's certainly worth it for Intel to continue working on the drivers, in the hope that by the time Battlemage arrives they know well enough how to make a solid driver.
I don't think Intel will cut its lossses unless it assumes that it will continue to remain behind on the hardware development front, which I think is something that Intel rarely assumes.
