Why intel redefined Boost and TDP

… could be related to their new comet lake desktop processors

ouch. Note : that’s not “overclocked” (well, one could argue it is, but does it count if intel is doing it from the factory?). That’s out of the box.

The thing is… if appropriately priced (e.g., cheaper than a 6950x was) and running at a sane clock speed I don’t think Comet Lake would actually be that bad. Sure it won’t be a Ryzen 3900x or 3950x, but for OEM use it doesn’t have to be.

All they’re going to achieve by ramping up power consumption to 200+ watts real world is going to be making life harder to OEMs.

Or maybe they don’t even care and its a total paper launch.

Comments? Do you think intel would be better off being realistic about the factory clocks on these things, or are they better off pushing things as far as they can physically go, and to hell with the power draw (and let OEMs figure it out)?

4 Likes

My humble guess is that they’ve found their architecture to be performing just as 14nm+++++++ has been performing or just above it, matching AMD. So they decided to crank that thing up to 12 to squeeze as much as possible out of it without any regard to efficiency. Also the quality of the silicon is of big influence due to their monolithic approach so they need to push every chip really hard to avoid silicon lottery variance influencing the performance.
What will be really interesting to see is if a chip that uses up to 224W can be overclocked with any success without exotic cooling.

At 224 watts you’re already looking at significantly better than intel box cooler… :smiley:

I mean my ryzen 2700x will push the corsair h115i with 280mm rad i have without boosting it any further. And that’s a 95 watt cpu if I recall, that only consumes something like 150 or less when boosting.

I wonder if there will be complaints from average users that they can’t maintain regular clocks if it is supposed to be 170W TDP nominal. That alone is a lot

1 Like

I wonder, under which circumstances (if ever) a regular user will ever see the top boosting freqs … Quite similar with the Ryzen 3000 “isue” of rarely if ever really catching their max boost freqs . Sure 5.3 Ghz is a catchy spec but , will the cpu ever reach that freq for more than a milisecond ? My older coffelake 8600k never ever reached the advertised its max single core/thread boost under normal settings and power limits . For that to happen , MCE should be enabled from the BIOS (which is essentially an OC).

1 Like

My 10-core 10900X at 5 GHz all-core consumes 400 W, more than a 64-core 3990X. No one who’s interested in this hobby cares about power consumption. A single 280 or 360 mm radiator is the limit beyond which there is no benefit - you’re limited by heat flux from the CPU and not cooling capability. You can drop a few degrees by increasing pump speed rather than fan speed, because that improves heat flux from the CPU.

Well, some do because basically power consumption = heat.

This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.