Extremetech: Intel Is Suddenly Very Concerned With ‘Real-World’ Benchmarking

He makes a few good points. But Intel or AMD, the big truth he doesn't realize he's making is that most people don't need a big computer. They might need nice peripherals, which is an argument lost in the budget storm of trying to buy the best computer. And the people who need nice peripherals like a big monitor, a nice keyboard and mouse, an easily accessed printer, or a good office chair rarely get them (think 'office administrator').

When I was programming for payroll we had to compile all of our code on the server. I didn't work in Visual Studio (I used UltraEdit, which I still like - columnar cut-and-paste, bitches!) so I simply didn't need a big computer. But I had lots of RAM and a $3,000 monitor.
 
That's a good article about Intel's presentation. Too bad Ryan Shrout found it prudent to sign his name on it [the presentation]. I guess one is not above writing BS with a corporate paycheck in hand.
 
Last edited:
Intels definition of "‘Real-World’ Benchmarking" is benchmarks that make their cpu's look good versus the competition.

Don't forget about a lack of patching for Intel vulnerabilities so the slowdown from the patches aren't there.
 
*Shrug* When Intel cannot win in the real field, they try to buy and BS their way through it, go Intel, you stay you.
 
Wow Intel... So i can expect at least 2 more years of getting your ass kicked then huh?

Considering how intel was making interesting talent hires last year, I'd say that seems accurate. Taking into account how long it takes for a product to be market ready.
 
Proudly brought to you by notebook and 2in1 tablets surveys !!

What, you don't run Photoshop and Blender on your 15w notebook? These are real-world benchmarks, man! :ROFLMAO:

Considering how intel was making interesting talent hires last year, I'd say that seems accurate. Taking into account how long it takes for a product to be market ready.

Intel's problem is that there isn't any low-hanging fruit left they can use to stand out, silicon just isn't going to scale much further. AMD has a head-start on chiplet design, TSMC has a process advantage, and what performance advantages Intel does have in their current chips seem to be tied to major security issues. The best they can do is catch up and try to differentiate themselves with better optimization for popular software.
 
Last edited:
Yeah, I LOLed when Intel said they are worried about real world benchmarks. Shows they are starting to get desperate, not sure how to take it yet, but seems like they're preemptively playing damage control.
 
I'm glad you back and can just write shit like this. I know it was short lived but it keeps us around.
 
Considering how intel was making interesting talent hires last year, I'd say that seems accurate. Taking into account how long it takes for a product to be market ready.
I can imagine the Jim Keller interview:
" I work when, and if I feel like it, salary is 900k with full management level benefits."
 
Maybe I'm becoming too negative in my years, but while I would welcome true "real world" benchmarking for CPUs, I'm guessing intel will use it as a marketing and obfuscation tool.

As for the Right Honourable Sir, I did read a while back he needed to leave due to a familiy sickness (hoping all is ok). In any case, he has an overflow of good will and respect here and in the industry I'm sure he could land somewhere very easily, [H]OCP or elswhere.
 
  • Like
Reactions: N4CR
like this
Haven't heard about that either. All the more of a waste to shut down hardocp then. It seems he had worked at intel a total of 1.5 months, 2 if I'm being generous :(
Yea, if I remembered the reason correctly, Kyle had family matters to tend to. I respect a person who put his family above his career. In anycase, Brent, Dan, and Paul now do review at thefpsreview.com.
 
Intels definition of "‘Real-World’ Benchmarking" is benchmarks that make their cpu's look good versus the competition.

Now remember a couple years back when some at AMD pointed out that most PC benchmark apps FAVORED Intel CPUs? Does anyone recall that most everyone laughed and called it sour grapes? The task of the day was to take turns pissing on AMD.

Well, well, well. How the tide has turned...................
 
Last edited:
funny how amd was in this bind back in the 90s. and did things like that with the k6
 
  • Like
Reactions: N4CR
like this
The K6 series were damn fine processors that were let down by poor motherboards. They outperformed comparable Intel chips in Integer ops while being a bunch less expensive. Intel held the crown in FP ops which were preferred by games like Quake/Quake II, that is true, but if your interests were in other game genres than the FPS, than the K6 series worked very well indeed. Except for those really shitty motherboards...

Actually, the most stable and compatible Socket 7 system I ever had was a K6-2/266 on an Intel TX Chipset-based motherboard. Runners up were the Tyan Super 7 Mainboards using VIA chipsets, but you had to constantly stay on top of VIA chipset driver releases.
 
The K6 series were damn fine processors that were let down by poor motherboards. They outperformed comparable Intel chips in Integer ops while being a bunch less expensive. Intel held the crown in FP ops which were preferred by games like Quake/Quake II, that is true, but if your interests were in other game genres than the FPS, than the K6 series worked very well indeed. Except for those really shitty motherboards...

Actually, the most stable and compatible Socket 7 system I ever had was a K6-2/266 on an Intel TX Chipset-based motherboard. Runners up were the Tyan Super 7 Mainboards using VIA chipsets, but you had to constantly stay on top of VIA chipset driver releases.
People say this because manganese back in the day were comparing the latter k6 to the older Pii with benches that catered to the k6.The k6-3 were out when the Piii were out, same for the latter the k6-2's
They were on par to the Pii in the real world. And this was if you were lucky and had a mobo with good l2 chips.
Not trying to bash the k6. I really liked it too, I had well over a dozen systems with of them over the years.
But the way the K6 was reported was kinda like if someone reviewed the 7/8th gen core to a bulldozer today.
 
Last edited:
The K6 was not the equal of the Pentium II in the real world. It was only close in some specific legacy applications because the P6 microarchitecture wasn't very good at anything 16 bit. Aside from that, not so much. Then there is all the Super-7 garbage motherboards and shitty third party chipsets you had to contend with.
 
Even then the Pii was older then most K6s, the latter and K6s that are worth having were sold at a time the Piii was out.
So they comparing a older CPU to a newer one. One of the many flaws back in the day when it came to the marketing of the k6.
 
Last edited:
Intel so MAD about cinebench benchmarks ... lol

maybe if they would have spent those years actually increasing processor performance substantially over the years instead of small iterations over and over
maybe no one would have cared if amd ran better in cinebench.

Intel dropped the ball and now they are paying for it, and its a good thing because AMD needed this . intel just needs to put on theit big boy pants get out there
and do what they best, and that balls to walls performance. insure thats its more than just a single digit increases and not get too greedy in the process and they will win back the enthusiast market again.
 
Intel so MAD about cinebench benchmarks ... lol

maybe if they would have spent those years actually increasing processor performance substantially over the years instead of small iterations over and over
maybe no one would have cared if amd ran better in cinebench.

Intel dropped the ball and now they are paying for it, and its a good thing because AMD needed this . intel just needs to put on theit big boy pants get out there
and do what they best, and that balls to walls performance. insure thats its more than just a single digit increases and not get too greedy in the process and they will win back the enthusiast market again.

The problem is that Intel can't do much of anything for a while. Everything designed was dependent on the 10nm process. Since the 10nm process was pushed back for years and effectively a failure it has set Intel back by years and it's unlikely Intel will have anything truly new or of note until they get the 7nm process working well.
 
The K6 series were damn fine processors that were let down by poor motherboards. They outperformed comparable Intel chips in Integer ops while being a bunch less expensive. Intel held the crown in FP ops which were preferred by games like Quake/Quake II, that is true, but if your interests were in other game genres than the FPS, than the K6 series worked very well indeed. Except for those really shitty motherboards...

Actually, the most stable and compatible Socket 7 system I ever had was a K6-2/266 on an Intel TX Chipset-based motherboard. Runners up were the Tyan Super 7 Mainboards using VIA chipsets, but you had to constantly stay on top of VIA chipset driver releases.
They were damn fine at dieing when overclocked a bit too much, without any voltage change!
After the first one died I had to check with another to be sure so I could warn people, yep it died too.
Expensive experiment.
 
I have an AMD A8... Still performs well in my 'real world' .. does that means its actually better than Intel's latest!! Mind blown. (Its actually feels faster than all the newer i5 and i whatever I use at work... Mind is REALLY blown at that one)
 
Back
Top