Tuesday, April 16
Daily News Stuff 16 April 2019
Chupakabura Edition
Chupakabura Edition
Tech News
- Intel's "35W" 9th generation CPUs have leaked. (AnandTech)
I say "35W" advisedly, because they only promise to run at 35W at their base clock speeds, which for the high-end 6 and 8 core parts range from 1.8 to 2.3GHz. If you actually get them into boost territory the power consumption will go straight off the chart.
- The EU Council has voted to approve the terrible horrible no good very bad EU Copyright Directive. (TechDirt)
No matter how bad a situation is, a well-meaning, focused, and properly funded government can find a way to make it worse.
- That done, the EU has moved on to its next terrible horrible no good very bad internet regulation. (TechDirt)
This time it's "terrorist" content, which as we noted previously, includes Project Gutenberg, the Grateful Dead, and C-SPAN.
- India has taken a breather from the Pewdiepie Wars to ban TikTok. (Tech Crunch)
Oh for fuck's sake guys.
- Journalist Carole Cadwalladr is a hyperventialting lunatic. (Tech Crunch)
Cool name tho'.
- Samsung is preparing for 5nm too. (WCCFTech)
Unless they aren't.
- Keybase... Lets you post signed social messages to confirm that you control accounts on various platforms. Kind of underwhelming.
- AMD's new Ryzen R1000 embedded APU family is really just a couple of low-end V1000 chips with a different name. (Serve the Home)
- Oh good. That's just what we needed. Adblock filter list providers can compromise your browser. (Bleeping Computer)
If you don't subscribe to any third-party filter lists, you're safe. This appears affect Adblock Plus, AdBlock, and uBlocker.
Appropriately Named Things Video of the Day
Disclaimer: Don't eat the yellow neutronium.
Posted by: Pixy Misa at
09:07 PM
| Comments (7)
| Add Comment
| Trackbacks (Suck)
Post contains 275 words, total size 3 kb.
1
Re: those 35W parts: they've got the T suffix, and while the boost speeds aren't out, the base clocks are, and they're about 1.6GHz slower than the regular versions (I only looked up a few of them.)
My work desktop is a Dell with an i7-8700. For most regular desktop use, it will maintain the boost clocks indefinitely. If I run something like Prime95, I can see that it will scrupulously limit itself to 65W, which results in it running about 200MHz below base. But for everyday use, like I said, it'll keep right around the listed boost. I would expect that these 35W parts will have lower boost clocks, probably nerfed about the same as the base clocks, and if you're not transcoding stuff or otherwise running them full-tilt, will probably stay near their boost speeds...but those boost speeds are probably gonna top out around 3.5GHz for the 9900T, and get progressively lower as you move down the stack.
So it's not just the turbo speed, but the duty cycle as well, that seems to control power usage.
My work desktop is a Dell with an i7-8700. For most regular desktop use, it will maintain the boost clocks indefinitely. If I run something like Prime95, I can see that it will scrupulously limit itself to 65W, which results in it running about 200MHz below base. But for everyday use, like I said, it'll keep right around the listed boost. I would expect that these 35W parts will have lower boost clocks, probably nerfed about the same as the base clocks, and if you're not transcoding stuff or otherwise running them full-tilt, will probably stay near their boost speeds...but those boost speeds are probably gonna top out around 3.5GHz for the 9900T, and get progressively lower as you move down the stack.
So it's not just the turbo speed, but the duty cycle as well, that seems to control power usage.
Posted by: Rick C at Wednesday, April 17 2019 03:29 AM (Iwkd4)
2
One thing I forgot to say: The Anandtech article says "This means that the 2.1 GHz base frequency of the eight core sixteen
thread Core i9-9900T might never be seen, and the power consumption of
the chip might be beyond 35W", but that depends on the BIOS--from experience, Dell locks their machines, at least the lower-tier ones like the Optiplex, to TDP. I would expect that to be the case here, too, because they also love putting these chips in SFF and micro-sized machines with, frankly, pitiful cooling, so they NEED to enforce TDP.
Posted by: Rick C at Wednesday, April 17 2019 03:31 AM (Iwkd4)
3
Waxing is not a cure for lycanthopy.
Posted by: Kristophr at Wednesday, April 17 2019 08:54 AM (9759K)
4
Once ran into a list for a hosts file level blocker that amid all the various malware sites it listed, also included a host of conservative blogs etc.
Posted by: Mauser at Wednesday, April 17 2019 12:18 PM (Ix1l6)
5
Rick - yes, it will depend very much on the BIOS. You can choose cool and slow or hot and fast, or more likely the vendor will choose it for you.
Posted by: Pixy Misa at Wednesday, April 17 2019 10:18 PM (PiXy!)
6
"more likely the vendor will choose it for you."
Yup. Buy one of these to put in a system you built yourself and I'd bet that the 35W limit won't be enforced, just like the higher limits on the existing chips aren't. But I would expect any big vendor, especially Dell/HP/Lenovo, to enforce them for regular home & business machines. And that is probably OK, but it may matter to the end user. (Now that I work at home, every once in a while I find myself doing work-related compute-intensive stuff on my home PC instead of the work one, because I have a higher clock speed.)
Yup. Buy one of these to put in a system you built yourself and I'd bet that the 35W limit won't be enforced, just like the higher limits on the existing chips aren't. But I would expect any big vendor, especially Dell/HP/Lenovo, to enforce them for regular home & business machines. And that is probably OK, but it may matter to the end user. (Now that I work at home, every once in a while I find myself doing work-related compute-intensive stuff on my home PC instead of the work one, because I have a higher clock speed.)
Posted by: Rick C at Thursday, April 18 2019 02:51 AM (Iwkd4)
7
Techspot has an article up claiming the max boost for the 9900T is 4.4GHz, the 9600T is 3.9, and so on, so we're looking at a loss of 6-700MHz, which I think is roughly consistent with prior generation T-class processors.
Posted by: Rick C at Thursday, April 18 2019 02:55 AM (Iwkd4)
54kb generated in CPU 1.0612, elapsed 1.0219 seconds.
58 queries taking 0.7187 seconds, 357 records returned.
Powered by Minx 1.1.6c-pink.
58 queries taking 0.7187 seconds, 357 records returned.
Powered by Minx 1.1.6c-pink.