So here is this monstrosity.

Once again, freezing up our loop is at 1.6.

See, because I am going to compete with myself.

Tired of all the extra cables in your system, Proctor design has you covered with their new eye on serious power supplies.

Available in a variety of sizes and wattage is the new fully modular ion plus syriza power supplies have what you need for clutter free reliability to learn more.

Click the link below.

So we learned a lot when I did the over clocking thing with Steve and kingpin.

Att.

The VJ anniversary bash thing will be over clock for like eight hours only to be beaten by Kingpin in 10 minutes.

Anyway, I remember what my CPU settings were.

And the problem was last time when we beat Steve with this set up.

I didn't have I didn't have a lot of time to dial in the CPU, so we did that kind of a crude, I think five points O r.

51 I remember.

I got to go back and watch my own video.

I remember what the settings were.

I didn't save any profiles because I'm dumb This is not a tutorial, though.

What I'm trying to do right here is I want to see if I can improve upon our previous score.

But I don't even know where we are right now in the Hall of Fame.

So let me see.

We're talking about specifically two times, Uh, grab your two ex graphics card or s Ally tighten our T X cards.

And I don't remember where we ended up like last time.

We were number five, but the gap between number five and floor, it's like 700 points.

That's not gonna happen today.

That's okay.

I want to see if I can be our 1 17,062 Steve has not fired back.

I don't think he's planning on it, but that's okay.

Let's see where we can end up a computer.

The computer restarted unexpectedly or encountered in unexpected airs Wednesday.

Window installation cannot proceed to install windows quick.

Okay, to restart the computer and then we start the installation.

So I guess a reinstalling windows potentially having to get more ice by the time we're done with this windows work perfectly fine.

Last time.

Shut it down.

Through the way saying, Dr Same motherboard.

Same.

See?

Few real windows installed.

That's good.

This is what we have to contend with here.

Um 17.

162 I just want to If I could beat that based on what I learned of Kingpin and the memory over clocking when it comes to the r T X stuff.

I wasn't pushing the memory nearly far enough, so I think we might be able to actually get an increase in our Jeep.

You score.

But I need to get our CPU score to at least match what we were.

And that was, uh, 17 605 I'm gonna go ahead and just run one right now.

Um, so I could see where I see Pugh score is all I've done so far.

When it comes to the CPU is our 31 75 x is over clock to five gigs.

All cores, chilled water.

The loop is sitting right here, a 2.7 c.

But anyway, what?

Um, what we got here is the fact that we lost a lot of our ice.

If there's still ice in here, then the water still hasn't absorbed all of the energy, which means we're still all right.

We don't have any salt in there.

I've never tried salt, but I look at it this way, though.

The couple of degrees I might be able to get a reduction.

And temperature is not the make or break difference of the scores that we're getting.

To get better scores, we'd have to go much lower than zero.

See, So I'm just going to a steep you test right now.

What I started to say is five gigahertz.

All core on the CPU, 1.3 volts.

We've got the Ram running at 3400.

Anything higher than that, We tend to start getting weird errors and stuff.

We'll play around with it.

1.35 volts on the Ram, not ex MP timings.

The timings are very loose right now.

One things we learned to the competition when Steven are messing around is that tight timings heavily benefit the score in times by extreme.

So what I want to see right now, it's just where we are with our CPU set to these very conservative settings.

OK, so 17 068 where the average of 20.5.

So again, if we look at our Hall of Fame results, I think that this was down, like, 17 or something like that.

So you could see that we definitely have so room to make up for come out in the go area.

So try to think I'm gonna go straight.

The 51 go 32 on our cash ratio without changing memory.

I don't think I got 51 toe work last time.

I think I think we crashed or we saw diminish return.

The crazy part about this is it's only been about three months since we did this last time.

Did I cannot remember anything that I did?

I cannot remember any of the settings.

I cannot remember the scores I care.

Like I'm doing it again for the first time.

So I want to see where I score came up to.

It was like 17,060 something.

And by adding another 100 megahertz, I'm curious as to what the score is gonna do.

We do know the CPU score is very heavy towards Ram.

Speed ran speed.

Now three confuses Ramstein.

Do I lost?

Du hast mish es.

We're seeing 17 milliseconds there for a second.

Are corer package tempers at 51.

52 crash.

I don't think we got 51 to ever run last time, if I remember correctly.

But I'm gonna try the brute force method by giving it a tiny bit more voltage because we have lots of CPU temperature headroom.

In that test, I wouldn't exactly fire up prime 95.

You know what I mean?

Because it would take prime 0.95 seconds to blow up 1.35 if I changed more than one thing at a time.

I won't know if it was bolted your cash that fix the problem.

So then if I brought the cash down to 30 and it was the voltage all along that could be losing out on 200 megahertz cash, which makes a big difference.

Yeah, I don't think five one's gonna work, which really sucks.

Dang, it blocked up reading the reset button.

So I got us a 1 17,037 100 CPU by dropping and back down to 5.0 gigahertz.

But going up to a 32 multiplayer on the cash, so I think I'm gonna try bumping up Multiplier are the cash as far as I can before I start playing with the memory.

It's important member to this memory is still running, running non ex MP timing.

So they're very loose.

Lots of things we can play with there.

So what I did in between right now because 5.1 wasn't gonna happen.

We left the cash in 32 but I also tightened up.

The timing's a little bit I brought down.

I care which one it is.

It was like a 38 right down to a 30.

And then I also increased the refresh time, and then I made it.

So it goes through 32 gigs of Ram before it swap, sir, or releases and renews starts writing RAM again.

So we're trying to do is just speed up the memory.

I also brought down the core only 10 megahertz, but I brought up the memory 100.

So we're fairly close to what we were last time.

We got our 17,000 whatever score that got us Number five on the Hall of Fame.

Now we're not gonna get number four.

That's not gonna happen.

There's, like 700 or 800 points between us or something like that and are at least 600 points and that's gonna take to get 600 more points, would take a core clock that we cannot get on ice water.

Nor can we get on a reference Titan PCB.

So now we're going to do C p.

U right here, which should validate with no problem.

But this is gonna be the first time we're testing those RAM settings.

These were some of the things I didn't get a chance to really do last time.

Interestingly enough to you're seeing this RGB Royal ram.

Typically, RGB ram is terrible for over clocking because the lights and the controller in there take power.

And when you over clock the ram to its limit, every military matters.

But when I asked the skill to send us more rams that we could test, they sent us this when I was like, Hey, I need something fast like last time.

The last time they sent us had no lights in it whatsoever.

They sent us this so far, it seems like it's doing OK because it's going faster than the last set that we had, which was the course there dominator But then again, that was only 3000 megahertz dims.

He's air 3600 so already right off the bat there, faster out of the box.

So what that means is they are not going to be pushing themselves nearly.

It's hard to try and achieve the score that we were with the course.

A Ram.

Oh yeah, 16 93 twos were like 200 points behind.

Ironically enough, if you look at our C P score, we're almost where we were, but we're not where we were.

That's a problem.

Problem benefits.

Right now, our good news.

We had to be the 17 605 were at 17 6 21 on our CPU score.

Um, I have not made the timing's as tight as it could be.

Yet I've upped it to 3800 megahertz, which is nowhere near where the Dominator could get it.

Like I said, I got 3400 or so Last time, sir.

400 extra megahertz timings air, still more relaxed than they were with the course there because I've got to start wide and dial everything in this part of the tuning that takes forever.

But now we know.

We've got to see if you back up to where it was when I'm gonna do right now, because I'm curious by, um, doing some testing here with our GPU.

I pulled our core back a little bit, but I brought our memory up to 1200 1 of the things Kingpin taught me was Look, if you start partnering with the memory in between, you get all kinds of weirdness.

If you just go to the max, a lot of times you kind of you shoot past this error territory and go right to wear, then suddenly starts working again, which is weird.

But that just seems so counter intuitive to me.

But I guess has everything to do with ratios and math.

And I'm not good at math.

We're not gaining anything.

Look, we got even a little bit more CPU, but we're not RG Pew is not cooperating.

So what I'm doing now is I'm wanting to see what's affecting our score.

Most right now want to see if its memory and I want to see or if its course red does.

I pulled the core back even more and added more memory, so I just sort of one extreme right now.

Way more memory, A bit less corps, because I want to see if we see the problem exacerbated.

If it does, then we'll know that we need to go the other direction.

So this is with 1300 megahertz, uh, over clock to the memory and you could see right now R F P.

S is even more stable than it was before, Lou, this It's only bouncing one or two.

F.

P.

S right there.

Extremely smooth, No artifacts.

So it looks like the artifacts that we were seeing were actually more core related than memory.

So now we're going to try is leaving the corps worth that and pushing the memory as far as I can.

So these air sometimes you've got to kind of go extreme, like you've got a test one direction, not the band.

And then you've got to swap it like 1 80 then measure what happened.

And then if you see a better improvement going one direction, then you start to go from you know you go outward from there.

There's no point if you see no improvement in trying like hammering Mork or Mork or more.

Four more.

And you're not seeing improvement.

Stop doing that and try something else.

And that's exactly what we're doing here.

Crazy.

We're not gaining anything.

See all that artifact in its core, get the f b s.

I don't know freaks out?

Uh, no.

Yeah, it's what we do right now is I'm just gonna go ahead and shut the system down.

Let's do a power cycle, See what happens.

Because that wasn't good way.

Just got the highest score yet.

17 679 It's only about 70 points higher than our best 74 points.

Whatever.

That's not gonna change our score.

A goal that just tells me I got the CPU back to where it's gonna be.

I don't think we're gonna ever get any more out of this CPU at 3.5 seed.

So what I kind of want to do right now?

Because I sort of just want to use heaven once again.

And I want to sort of monitor when our clock behavior is that showing 2190 with a plus 2 20 plus 200 breathing down 2175 I think Want to try it with 1 85 on our core.

So yeah, it's bouncing all around, which is normal, but GT two all over the place so I can't match my old score.

It could have everything to do with the fact that my cards could be in the wrong order community.

I don't know which card was the better card last time, which is stupid.

I should be able to tell when is a clear block one of the black bloc?

Everybody easy to keep track of that.

I'm allergic to failure.