Or the CPU was launched before it was fully ready and needs a microcode update? Or Hallock is doing PR damage control? Or he's mainly talking about all of the weird issues (like instability in 24'H2 and APO not on be default)?
Im not making any claims to the accuracy of Hallock's statements regarding a performance uplift incoming with an update. I'm just simply pointing out that:
1) "performance is not where we would like to see it and are working on a patch" and
2) "the performance 3rd party reviewers see and what we see align"
Are **not** mutually exclusive statements. Both *can* be true.
This is the crux of what disappoints me regarding Intel's handling of this problem. All we know, all they will divulge is benchmark results were "unexpected" and we're going to fix it?
Fix what? Why? Maybe the benchmarks were higher than they expected? Who knows. People are happy because Intel actually popped their head up and said something. OMG they listened to us, the little people. They are only responding because this is not good for sales in what could be a big way. But still, it's Intel and they generally aren't forthcoming (read honest).
One of the only times I remember some honesty was the interview with Raja Kodure regarding Arc. He said they didn't expect the driver issues they were having because they had so many years creating iGPU's, what could go wrong? They'd be doing drivers forever.
Then he admitted when you throw more hardware into the GPU all sorts of bottlenecks in the software begin to pop up and this is something that pretty much hit them from out of the blue and it was a big problem that required and still requires a lot of hours to work through. He also admitted that some of the design hardware-wise was stronger than it needed to be in areas and not as strong in others and this also came back to the difficulty of creating a balanced design that is 10x or 100x more powerful than what you had been building.
Now to me this all made perfect sense and Intel was and still is cranking out driver updates that have done wonders to make Arc more performant and more stable. 8 months ago Arc and Topaz AI apps was dicey, today it actually works quite well. It does not hurt to try and obviscate issues. Most people, probably 90% only need to hear a fix is coming and they move on with their lives. But others, like me, get frustrated with the lack of transparency when it comes to a product I may have already spent my money on or are considering for purchase.
I like Intel. I want them to succeed. I also like AMD and want them to succeed. They both have done amazing things and I'm sure they will continue to do so. I'm only saying more honestly, not less, will lead to greater brand alligience and trust from the public, not less, which has been and continues to be their business practice. This is my opinion so it is worth exactly nothing.
Now back to our regularly scheduled program. I will wait with baited breath for the resolution of ARL.
Specifically, I have three questions I'd like to eventually have the answer.
1. Was the press release performance data what was expected of the final product and is that what or not what we saw in reviewers benchmark data?
2. If that is what we saw then why is Intel talking about Windows and BIOS updates to fix the "unexpected" results?
3. If that is not what happened, meaning Intel press release benchmarks and reviewer benches were in alignment, then why is there any talk of a fix? Where did that come from because it couldn't have been from reviewers benchmark data if that data supported the press release data and vice-versa?
3. Please provide insights into the BIOS and Windows changes. I mean you have divluged so many details regarding the architecure how could it be harmful to your IP to tell us about the software bottlenecks you are fixing?
Something is not striking me as logical thus far.
AMD was different. Some benches were low, AMD agreed, told us why, and fixed it. End of story.