Intel gives up on benchmarks

They loved them when they were winning, then changed to "real world benchmarks" they could still win at.... Now they want to stop using them until they are on top again. Makes sense from their point of view... If I was losing a basketball match I'd want to stop counting the score and base it off something I could win, like brightest jersey color! Is this surprising?. Always use a met if you can win at if you're trying to sell something. AMD isn't running around with gaming benchmarks it loses at to prove they caught Intel.... They use ones they do better at and apps they do good in.
 
Benchmarks are only as useful as the comparisons between products are relevant to the users buying the products. It is fairly straightforward show a comparison between benchmarks for performance, power draw, speed, etc., that are (relatively) easy to collect data for and report on. Yeah, synthetic benchmarks have their limitations but when more and more benchmarking is migrating to real-world usage such as in-game demo performance, productivity benchmarking, etc. that excuse gets watered down.
There are other things that are harder to quantify such as ease of use, upgradability, hardware interoperability, modularity, manufacturing benefits, etc. that people may pay more for.

Now sure, running your production plants off of "green" energy and taking steps to reduce waste and toxicity of production processes are great, but if your primary business is to sell performance per dollar to consumers it will be a dry conversation.
So in my opinion his paragraph response from the article is fluffed to make things sound better than they are - but you can expect no less from a CEO that is trying to protect the brand.
 
Fun quotes from Intel CEO Bob Swan

https://www.guru3d.com/news-story/v...computex-did-intel-give-up-on-benchmarks.html


"we have to stop worrying about things as simple as the benchmarks of processors"

Seems like a desperate move to me. What do you guys think? Is onto something or full of something?
1591203627057.png


1591203661654.png
 
Bob Swan: "we have to stop worrying about things as simple as the benchmarks of processors"
Everyone else: "we have to stop worrying about things as simple as Intel processors FTFY"
 
Seems to be that Intel is in the market to sell more Intel CPU's and their marketing has shifted away from performance and power and IPC, to... Look at all the Intel CPU's we work with. You don't want to chance that do you?

And honestly that speaks to 2 kinds of customers.

1. Enterprise customers who don't want to risk an relative unknown.

2. Non technology savvy buyers who just want a laptop they can use to do what they will do. (or desktop but that is a waning market outside of Build your own if the advice I've been giving friends and family is any reflection.)

So everyone else... Engineers like me on then enterprise side will be working to see IF AMD's solution is better. Because honestly the cost of my servers is more tied up in Licensing than it is in hardware costs. And developers are making it harder and harder to get approval for big multi core servers outside of a few niche cases. Really in the end it comes down to 1 kind of server that the business wants lots of cores from a hardware/licensing perspective and that's Virtualization. But companies like Vmware (the big player in my opinion) is moving away from per socket licensing to something more core based. Just a few months ago they switched over to 1 license per 32 cores in a socket. So if you have 34 cores per socket you are paying for 2 licenses per socket up to 64. So it's either go whole hog and get 64 or in the future 128 core CPU's, or ride the licensing razors edge and go with 32 cores a socket and hope they don't tighten the reigns more.
 
I think 14nm has taken Intel as far as is physically possible without getting into 500+ watt TDP ranges.
Outside of vendor lock-in or edge-case gaming, I see little reason to invest in Intel CPUs in this generation, especially with their lackluster chipsets thus far.

This is just sad, and if Intel doesn't get its act together soon, due to lackluster competition from Intel, AMD is going to stagnate throughout the 2020s just like Intel did throughout the 2010s.
If that happens, it will be the death of x86-64 by the end of the decade in favour of ARM64.
 
90% of their sales go to office computers and laptops that will never get above 40% CPU load as is and those rare times it does will because it is installing something large or unpacking a zip file. For the past 5 years or so Intel's CPU's have been more than fast enough for what they mostly do it was only crapy HDD's that were holding them back. But now that Everything is a wep app, or an RDP connection the physical device matters very little. I could downgrade all but a small handfull of my users from an i7 to a Celeron and as long as I had a good M.2 in that system they would see it as an upgrade.
 
90% of their sales go to office computers and laptops that will never get above 40% CPU load as is and those rare times it does will because it is installing something large or unpacking a zip file. For the past 5 years or so Intel's CPU's have been more than fast enough for what they mostly do it was only crapy HDD's that were holding them back. But now that Everything is a wep app, or an RDP connection the physical device matters very little. I could downgrade all but a small handfull of my users from an i7 to a Celeron and as long as I had a good M.2 in that system they would see it as an upgrade.
A lot of truth in this... but that super high TDP!!! :D
 
Bwahahhahaha. What a laugh. You know your product is complete and utter shit when the best line your CEO can come up with is essentially translated as, "don't worry about our products performance".

Dead dickwad Bob, computing performance is the whole point of CPUs.
 
Back
Top