this post was submitted on 30 Oct 2023
1 points (100.0% liked)

Hardware

47 readers
1 users here now

A place for quality hardware news, reviews, and intelligent discussion.

founded 1 year ago
MODERATORS
top 16 comments
sorted by: hot top controversial new old
[–] [email protected] 1 points 1 year ago (1 children)

They skipped an interesting range around 190W where you barely lose anything in performance

[–] [email protected] 1 points 1 year ago

Part of it is operating at a more efficient point on the VF curve. The other part is that Intel parts increase voltages by ~2.5mV every for every degree C closer to TjMax. A part at 95C@253W has a 50mV worse VF curve than a part at 75C@190W which can be a substantial efficiency difference.

[–] [email protected] 1 points 1 year ago (2 children)

Less than 5% performance loss at 1080p from 253W to 95W while only 12% for applications. Factory overlocked for performance crown nonsense makes these extremely inefficient.

[–] [email protected] 1 points 1 year ago

That is honestly fantastic. I guess I can get a 14700k and comfortably combine it with a Peerless Assassin. Great bang-for-buck.

Now if only there was a cheap z790 option for 14th gen :/

[–] [email protected] 1 points 1 year ago

And everyone in the AMD subs downvoted me en masses for saying Raptor Lake was actually pretty damn efficient when not chasing peak performance. My 13700k sips a mere 50w while playing a CPU intensive game at 4k120.

People just assume it'll always use as much power as when it's benchmarked and can't be told otherwise, even when they don't own one themselves. Still, these results are pretty amazing even to me. I'm curious what I can do with my 13700k now at 95w with an undervolt.

[–] [email protected] 1 points 1 year ago

U and T chips rule this chart https://www.cpubenchmark.net/power_performance.html and that's what this test shows too: the 35W version obliterates everything on the desktop efficiency wise. Ironically, it's even more efficient than the E core only N100.

It's quite interesting to see this in light of https://benchmark.chaos.com/v5/vray?index=1&ordering=desc&by=median&my-scores-only=false -- on server platforms AMD is way more performant. Perhaps Intel's core doesn't scale as well up as it does down?

[–] [email protected] 1 points 1 year ago (1 children)

This is hardly new, 13th gen (12th too?) behaved similarly, it was shown on multiple occasions that they can be just as efficient as Ryzens, that Ryzen's power efficiency doesn't translate into good thermals, etc...it didn't silence the AMD cultists on the internet and sadly this won't either.

[–] [email protected] 1 points 1 year ago (1 children)
[–] [email protected] 1 points 1 year ago (1 children)

Anandtechs testing is completely wrong because they used TDP on Ryzen and PL2 on Intel. PL2 is the actual power limit, but TDP isn't. PPT on AMD is 1.35x TDP and is the actual power limit, which is why the measured power is much higher than the limit they set.

It's absurd that they've never issued a proper correction for an article that has the AMD CPUs allowed to draw 35% more power at every comparison point. It's been misleading people for years.

[–] [email protected] 1 points 1 year ago

No you wrong, read article they changet PPT for Ryzen it is power limit like PL2 on Intel.

[–] [email protected] 1 points 1 year ago

I love powerlimiting

Imagine overclocking in 2023

[–] [email protected] 1 points 1 year ago (1 children)

For point of reference. A 13700K tested using R23.

Watt Score Score/Watt
15 4200 280
20 6800 340
25 9200 368
30 11200 373
35 13000 371
40 14500 363
45 15700 349
50 16700 334
55 17600 320
60 18500 308
65 20279 312
80 22391 280
100 24776 248
115 25667 223
125 26102 209
150 27384 183
200 29582 148
[–] [email protected] 1 points 1 year ago

This is what I have been looking for! How does this compare to Macs?

[–] [email protected] 1 points 1 year ago

What the holy shit man. A 14900k does THAT good when severely power limited and undervolted? I wonder if my 13700k will be similar. I know if you power limit it to 200w it only costs a couple percent of performance. Wonder what happens if I go down to 125w? Or even 95w? This requires further testing....v

[–] [email protected] 1 points 1 year ago

I just compared this test to another done with the 13900k at low power levels and it seems like the 13900k performs better at 125w and below, or am I missing sth? Curious how the 13900ks is at 125 or 175w. Probably even more efficient due to binning.

[–] [email protected] 1 points 1 year ago

Might be a good place to ask, but how do you guys get around XTU not working when you have secure boot on? Is the only way to powerlimit/undervolt through BIOS?