You being a bible student might offhandedly wonder which is real.
Whichever passes your teacher's test?
Now switch over to BIOS pontifications, ignoring the truly ignorant
and the apostates, the mad-people (people who've gone mad reading too many fragments of thoughts)
It's not exactly ex-cathedra but it sounds well-thought-out, so it goes in this-here entry.
I left them on "auto" and it changes like moods, but right this second it's 8.4
8 because (I suppose) it's 1/2 of 16, and "4" because it's being careful.
https://github.com/integralfx/MemTestHelper/blob/master/DDR4%20OC%20Guide.md What is TWRPRE/TWRPDEN in AMD's language??? (huh??) "Twr=twrpre(yada bleagh) |
If you went to that page, or a similar one, you'd surmise (OK *I* surmise) that some settings override others, and your petty attempts to modify them, your puny insignificant...(well anyway)
My TWR on an AMD board will not go to 28, it's stuck at "26"
This isn't unheard of, because changing "Gear down mode" to "enabled" is like hiring a pontificator to preach at your computer.
O...K.
Just, um,
ignore me...or something.
Edited-in Rant: No one talks english, they raise it to another level and it's up to YOU to decipher what exactly is meant at page 21-paragraph-3 (know thy reference, study to show thyself approved)
If you have a 3950, use 4 single-sided DIMMS, Not Two double-sided. (4 double-sided will lead to outer darkness and tooth gnashing) |
And we don't even know why guys write this stuff, was it for a class, a degree? How far could you get pasting stuff and changing a couple words? The PDF in the link sounds a LOT like something proposed, not invented yet.
Strange.
Well, maybe they were well connected.
Useful sites are really hard to find, But It's a formula or a set of formulas we're looking for,
Amazing, if true (it isn't) |
Tras=
TRC=
TRFC=
Tfaw=
TRRD_(S,L)=
No one reads this but me, and I can't see this while updating BIOS,
But...
Maybe some of it will stick in my head, or I'll remember to take pictures.
Waterboard it, render it until death, then back off (whadda crock) |
cl (or 2*cl) Do they even listen to themselves? I think they meant RCD_r+CL+2, but I'm translating mental-hospital jargon to english,,,I'll come back to this. |
What, you thought I was kidding? (What am I, a clown for your amusement??) Check the sources, don't just bow because the pontificator says "wikipedia" |
But...if 2*RTP is TWR, and TWR is stuck on 26,
uh,
1/2 TWR (26) is 13 (as opposed to 14), so RTP should be 13.
I think.
Except, 13 is an odd number, (also "13")
Edit: TRTP is *also* stuck, at 14.
https://www.techpowerup.com/review/amd-ryzen-memory-tweaking-overclocking-guide/6.html |
If the smart guy is correct and did not stammer, TRTP should be 6, TRAS 32, TWR 12.
O...k, here goes:
But then he gets prosaic, and I get lost, slightly:
(I mean, once something works, do you care?) But here:
He contradicts his own self and brings in variables from the Twilight zone.
Let's calculate:
TRAS=16+4+12
20+12=32.
O.
(nevermind...)
This inscrutable thingy is saying (if you squint) that WR and RTP are related |
And I shall bend into windmills now for an hour and come back to gloat, or sheepishly erase this paragraph.
Yes, apparently some new framistan related to DDR refreshing is uh,, o nevermind, forget I ever mentioned it; tons of links to bombastic blogs would be in order.
But anyway, that was no typo! It worked.
(For now)
This is greek to me, but maybe you read greek, How do I know. |
This many faceted rant is saying MY old duffer DDR4 can run faster now,
either through new technology or tech that is a compromise for existing stuff
by really smart (but very quiet) engineers working at DDR factories.
Or am I to believe that a proposal written around 2016 is already in my system?
No, I refuse to believe it (but I'm just, ya know, mumbling)
Maybe they added a cap or a resistor, but I doubt they added the rube-Goldberg thingy ChargeCache yet (into my PC)
Snipe spotted at http://www.ddr-phy.org/forum/topics/information-about-tphy-rdlat?overrideMobileRedirect=1 |
A Russian site has a file called "Zen timings", part of a train of thought from
That is mostly unreadable. Geniuses looking cool (" 'sup, dog")
Being a total-snipe hunter, I will now unzip the file and run it...if I'm not back by tomorrow, uhm.
BTW if someone totally *gave* you $50, uh...?
Well anyway, another snipe-hunter thing to do, take the money and run to the drug store...
I don't know.
Another flibbertigibbet from Amazon?? ( I stand corrected, it's actually "Thingamabob" or "gewgaw")
Edit: I went super-Caligula on my ass, bought three biggie Peach-pies, ice cream, Avocados and Sour Cream, only the best.
My butt will go "Ptui" eventually but...
So...we're good, right? (Right? oGawd) |
"Let's talk DRAM!" from AMD has a "Privileges Denied" warning, you have to be *worthy* to receive the drops of wisdom from their lips!!
And what do the (stolen drips from the) lips say??
Speak, o lips which hath no tongue:
"So", you ask eagerly, "does 12 work?"
I don't know, I feel so sheepish.
Come back in an hour (Aida is NOT impressed, readings, same)
Bleah...prime95 is next.
Pointless if there ever was, a waste of an hour.
But yeah, it worked.
"[Worker #14 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
The above "prime95" picture shows how rocky my knowledge is, of the whole tweaking process...
"[Worker #14 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #11 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #9 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #10 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #13 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #12 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #16 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #15 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #7 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #5 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #8 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #1 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #6 Jun 27 10:45] Torture Test completed 26 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #4 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #2 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings.
[Worker #3 Jun 27 10:45] Torture Test completed 22 tests in 1 hour, 30 minutes - 0 errors, 0 warnings."
The above "prime95" picture shows how rocky my knowledge is, of the whole tweaking process...
*Some* CPU's had enough time to complete 26 tests, and some only had enough speed for 22 tests, and Why, I will never ever know.
All I do know is, they ought to make this easier.
One single guy pontificates 12 TWR,
and RAM manufacturers are oblivious or holding out for more money
("Now with Blang")
I'm not sure I ranted clearly today, and I'll forgive myself because the information is so opaque, the benchmarks, so expensive (and mostly, useless.)
If you already own a beloved benchmark, with dozens of scraps of paper with results from 2006, congratulations.
If, however your benchmark updates itself constantly, has a retail price that could buy groceries for a couple days, and you only just got it,
too bad.
I still have a benchmark (memorytest) somewhere, if the bugs haven't eaten it.
And I've used "Aida" long enough to know that a latency of 66 isn't slow, if 55MPH isn't slow.
63 is faster, 61 is enormously fast, etc.
But why pay 49.00 for a latency program? cheap bastards can't put in a delay so mouse-clicks will not affect the results.
"Passmark" seems less respected, is more for burning up the PC under warranty or before you take it home.
Are the Latency results so very far off????? OK maybe I need a free trial to compare aida and passmark. |
So I'm lost, bereft.
I have money, I just wouldn't spend it on such crap.
But as compensation for not having had it for years and without 5 or more personal scores from past PC's, there's this online site I can look at scores. (wait, didn't I do that?)
Too soon to tell, too tragic to joke.
But can I draw your attention to my rotting rug?
https://www.usatoday.com/story/news/investigations/2021/06/24/building-collapse-miami-structure-had-been-sinking-into-earth/7778631002/Too soon to tell, too tragic to joke.
But can I draw your attention to my rotting rug?
The latest spin on this story is that they've known for many years that the building was shifting.
And I want to quote a line from that article, all due respect
I wrote an entire entry on diabetes and the various points of view,
and then erased it, because it got too ranty,
too bombastic. (what I said about windows, before I erased it.
But right this second I want to go to the store to buy food, when I already have a refrigerator full of food.
Bland food.
What about ice cream?
What about salad?
Oooh, with vinegar!
Isn't there some (justifying) website on vinegar??
https://diabetesaction.org/article-vinegar, it is or is not helpful
(which goes to my "zillion bibles" rant.)
Bah,
You're all nuts right now, best to go when you're all sleeping it off.
Tough sitting here dreaming of my grocery-list.
No comments:
Post a Comment