themaniac
-
Posts
5,182 -
Joined
-
Last visited
.gif.c2d7c53e12a6c2a8f676e7abd35abf3d.thumb.gif.b82aaab771db565158104305fdaa6e37.gif)
.thumb.jpg.c24e50caf0f454236db021676ad49609.jpg)
Recent Profile Visitors
themaniac's Achievements
-
i have thoughts, dumb thoughts but thoughts none the less. thoughts of ideas, bad ideas, but thoughts of ideas none the less
SO THEREFORE BEHOLD MY TEN SECOND ARTIST RENDITION OF MY BAD IDEA THOUGHTS
THE PC SANDWHICH!!!!
and now i disappear for another few months
-
I definitely get off and on interested for tech these days, but seeing this bin has brought my interest back. Really tempted to ask IT if I can take a few things from it. So i can actually try learning more, no idea what exactly is in there, all i know is that i see some rack mountable things in it. Having never seen what a 1U, 2U, or 4U looks like im thinking they look like 2U.
i also blame this video for bringing the interest back as well
-
assuming these benchmarks are accurate i find it extremely interesting how removing hyperthreading improved multicore performance so much.
After reading it over after i finished writing it all I decided to spoiler it. This is your warning, read the mess of number comparing between intel 13th gen, arrow lake, and amd 7000 at your own risk
SpoilerI looked at other scores on cpu-z comparing similar tier amd vs intel and it makes me wonder if intels implementation of smt/hyperthreading is just really terrible.
For example the Ryzen 5 7600x vs the i5 13600kf - The i5 gets a multicore score of 5970 with 20 threads vs the r5's 5919 with 16 threads
so per thread its 298 for intel and 370 for amd
meanwhile single core scores are 823 for intel and 734 for amd
Now for the arrow lake i5 leak, its 850 for single core and 10907 for multicore or 779 per thread
Which brings up another thing how good are the new e cores? Because the leaks seem to show that the new i5 is 6p 8e cores.
Again assuming these numbers are all correct and we take them at face value. The P cores would make up 5106.6 of the score with the E cores taking the remaining 5800.5 score or a single core score of 725.06. Which is 85% the performance of the P cores.
Now lets do the same to the 7600x
single core score of 734, with a multicore score of 5919. The normal cores make up 4404 of the score while the 6 threads from smt make up the remaining 1515 or 252.5 per thread
Now we shouldnt expect extra threads alone to equal a full on core but it does make me wonder how much benefit is smt actually.
Does building smt into an arch hurt the overall performance even if you disable it in the bios vs building an arch without smt in the first place.
I would do the same for the 13th gen i5 as well but i really have no idea where or how i could separate p core from e core from smt threads. For all i know all the potential uplift in arrow lake is actually from intel improving the e cores significantly. double the performance or something
HOWEVER
I did more looking after writing all of this and found the 13600kf in the 20 thread count section at 9630. somehow completely missed that, the other score was in the 12 thread count section so.... time for more math that more than likely doesnt actually work for actual comparison but i shall do it anyways
Alright so with a multicore score of 9630 and a single of 823. the cores alone make up 4938 of the total with the threads and e cores making up the remaining 4692.
If we take the 12 thread score of 5970 into account, which im gonna assume is with e cores disabled/unused. So taking the two scores and subtracting that gives us 3660 which in theory should be the e cores score, giving us a single core score of 457.5.
Now for the threads, going back to the 12 thread score of 5970 now that we "know" this is without e cores that leaves us with 1032 for just the threads or 172 per smt thread
smt v smt | Amd v Intel, intel gets 172 vs amds 252.5 or intels smt is 68% as good as amd's smt
Now lets consolidate
AMD 7600x
Single Core: 734
Multi Core: 5919
SMT: 252.5
Intel 13600kf
Single P Core: 823
Single E Core: 457.5
Multi Core: 5970
SMT: 172
Intel Arrow Lake I5 Leak
Single P Core: 850
Single E Core: 725
Multi Core: 10907
So all in all, my dumb math which more than likely does not work at all in the real world puts Intel's smt implementation behind AMD's with Intels being 68% the performance of AMD's. While Intel's single core for 13th gen is better than AMD's and E cores 55% of 13th gen P cores and 62% of AMD's cores.
Then we compare with Arrow Lake leaks, which puts Intels E cores roughly equal to AMD's Zen 4(and probably Zen5 as well looking at how thats going) and P cores ahead of AMD. And the new E cores are a massive improvement with a 58% increase / old E cores being 63% the performance. And being 85% the performance of the P cores
I dont know exactly what i accomplished from this but its over now, enjoy your freedom from my stupidity
-
Cinebench has long been the best "jack-of-all-trades" benchmark in my book, with specific use cases being lost on it because they're, well, specific (e.g. the X3D chips & gaming). So until that (and the 245K itself) comes out, pour out the salt.
However, if this holds, what they really found is that the E-cores were being held back by the hyperthreading monumentally, which makes sense for reasons of heat and power draw.
-
-
-
well, made the swap to the 12600kf from the 8700k, only did a few tests, but its a pretty significant jump. However windows said fuck you and deactivated itself and refuses to reactivate.
Tried buying a key from g2a, didnt work, and now i gotta wait for the seller to reply and make excuses for why they wont refund.
-
4 hours ago, themaniac said:well, made the swap to the 12600kf from the 8700k, only did a few tests, but its a pretty significant jump. However windows said fuck you and deactivated itself and refuses to reactivate.
Tried buying a key from g2a, didnt work, and now i gotta wait for the seller to reply and make excuses for why they wont refund.
You should try hitting up Microsoft support and see if they can re-activate the old key.
I've had two of my keys re-activated before like that and both were OEM keys.
As for the G2A key, have you tried the phone activation method? That usually works for them.
-
9 hours ago, Cyberspirit said:
You should try hitting up Microsoft support and see if they can re-activate the old key.
I've had two of my keys re-activated before like that and both were OEM keys.
As for the G2A key, have you tried the phone activation method? That usually works for them.
should have tried microsoft support, but i was able to get the g2a key activated with the terminal.
-
-
looks like its time for a password change
Absolutely no way microsoft would say "someone else might have typed your email address by mistake"
-
For the low low price of driving a guy an hour and a half through minnesotan rush hour to microcenter, I got a decent little upgrade in return, a 12600kf, going from an 8700k so im excited to see how much that changes things. Sad part is that my current cooler didnt come with the mounting kit for lga 1700 and it wont arrive until the 1st of july....
-
Got a dash cam because it was on clearance for half the price($50 from the original $100) and it has a max of 32gb microsd card support. Why is that an actual thing? Why is there max storage support for sd cards? Is it a license fee issue? Different pinning/extra pins on different sized sd cards? Harder to program/code support for? Or just some bs?
-
So.... Why is it that my phone loves to send me notifications with articles that i days old?
Why is it that i got this today, 2 days late, so I actually thought it was real for a good while?
-
Shit, we're already a quarter of the way through the year. Thank you temporary font change for reminding me....
- TopHatProductions115, soldier_ph and Kilrah
-
1
-
1
-
1
-
TFW you visit a different location for your job and it makes all the issues with your location look like nothing.
I'm surprised that location hasnt been shutdown by the food safety people and osha yet.
-
so to give context to my previous status(completely forgot about it after changing tabs). My dad has a shed in the drive way thats right not being used for storage. I recently got the idea that I could turn it into a little workshop of sorts. So before I empty it out i need a place to move everything. So I thought about building a loft in larger shed my dad has. Materials cost money.
Then once I empty it I need to run power to it, but I also remembered I have a 3d printer and an older pc i can set up. So I'll also want to run internet to it. AND I live in wisconsin, so the winter gets quite cold(not this year for some reason) so I'll want to insulate it and heat it. Along with putting up sheeting for a bit of fire resistance. So more money.
MONEY IS THE PROBLEM OF EVERYTHING
IDEAS COMPOUNDING ON IDEAS ARE THE PROBLEM OF EVERYTHING
-
30 minutes ago, themaniac said:but I also remembered I have a 3d printer
How do you forget that you own a 3D Printer ? I mean a 3D Printer is the cool kid, unlike yestern-decades 2D Printer.
-
its such a cool kid that it got shamed into the closet and forgotten. But soon it shall come out of a closet with new founded confidence.
-
-
I idea, ideas that out idea my ideas so much that my ideas wish they never idea'd
-
first two went decent, went, but then the white bastard, THAT WHITE CABLE just had to ruin my success
its disgusting so dont look if you want to be safe
Spoiler
pretty sure the left cup is slightly louder as well because of it.
-
If I had to guess what happened, you didn't use additional flux and you had to heat it up so much the flux that's included in the tin (usually) just vaporized and the insulation melted
And the solder doesn't bead up nicely as well
Additional flux is very nice to have if you want clean solders
-
11 hours ago, Moonzy said:
If I had to guess what happened, you didn't use additional flux and you had to heat it up so much the flux that's included in the tin (usually) just vaporized and the insulation melted
And the solder doesn't bead up nicely as well
Additional flux is very nice to have if you want clean solders
I had my soldering station set to 400F, i had looked up the temp for the 60/40 solder i used, what i found said 400F was a good temp, so maybe it was lack of flux, since i just reused the solder that were already on the pad
-
13 hours ago, themaniac said:
I had my soldering station set to 400F, i had looked up the temp for the 60/40 solder i used, what i found said 400F was a good temp, so maybe it was lack of flux, since i just reused the solder that were already on the pad
if you dont wanna buy additional flux, just remove the old tin from the wires, then heat up the area before applying new tin, it should melt the instant it touched the things you want to connect.
this will minimise the amount of time the flux included in the tin have to vaporize
-
so my cats had chewed through my headphone cable a few times last year, i've put off replacing it this time for a few months at this point. Decided to get around to doing it now. And now i remember why i put it off.
There's very little slack in the wires to move the driver into an easier position to solder, plus due to how headphones are and like to clamp back together, its tough to get it in a position to both have the drive be somewhat possible to solder AND not clamp back together.
Now i want to stab myself with my soldering iron
this is the best i can get and im still gonna hate every second of it
Spoiler
-
seems like its not possible to use an nvme based m.2 in a sata adapter, was thinking of putting one in my laptop to give it a little boost. Since the drives are just sitting collecting dust. But guess not
.png)
