Forums first comp: CPUz Benchmark

[Avacado]

1 hour ago, Storm-Chaser said:

Anyway I can link my pre-existing storm-chaser account, it's forcing me to sign in with google or facebook which I don't want. I dont want to create a new HWBOT account if at all possible.

You don't have to create a new account if you already have one. Just need to change teams. Or just join EHW if you have never joined a team in the past. Go to your "My Account" Page. Should also be possible to change/join from the "My Team" Page. 

 

Edited April 4, 2022 by Avacado

[Storm-Chaser]

1 minute ago, Avacado said:

You don't have to create a new account if you already have one. Just need to change teams. Or just join EHW if you have never joined a team in the past. 

Ok thanks. Btw my 9600k rig had been repaired and is now out of the hospital. Praying for cold weather. Might actually be able to give mr fox a run for his money on his 5723 MHz run! I think the liquid metal is key if I want to be competitive here. The few runs I made last night failed half way through the test. But I just need to ride that knife edge of stability.

[J7SC_Orion]

5 hours ago, Storm-Chaser said:

Anyway I can link my pre-existing storm-chaser account, it's forcing me to sign in with google or facebook which I don't want. I dont want to create a new HWBOT account if at all possible.

 

...yeah, re. creating a new account, you have to be careful with that re. HWBot rules, at least that was the case years ago...worth checking. It used to be the case that you could transfer your scores 'once' into a new-name account. This is different from changing teams in your existing account / name and taking the points with you.

[J7SC_Orion]

@Bastiaan_NL , @Avacado , @Mr. Fox

 

...for your perusal...

...or Raptor Lake  ?

 

 

[ArchStanton]

Interesting video/report.  I spent an embarrassing amount of time trying to figure out what RGBismus is (basically RGBism in English, a "catchphrase" of some other German TechTuber).

[Mr. Fox]

2 hours ago, J7SC_Orion said:

@Bastiaan_NL , @Avacado , @Mr. Fox

 

...for your perusal...

...or Raptor Lake  ?

 

 

Too funny, bro. Great minds do think alike. I posted that over here earlier today.

 

I did not really expect the KS to be a different product as much as a binned sample of an existing product. So, there is some basis for disappointment. However, taking the glass is half full view, buying a KS might be the best way to avoid being dealt a tragic ticket in the silicon lottery. That's not a bad thing per se, but paying $200 extra for something that isn't as good or better than a superior 12900K silicon sample is where the basis for disappointment comes in. It's more like damage control  or loss prevention than an upgrade. You have to decide whether it is better to roll the dice on a K and hope for a win, or pay extra for something that is, at least, pretty good... with potential to be better than pretty good. If the chips are now being culled in favor of KS branding, then the chances of winning with a K might be drying up as old stock gets sold and replaced with stock representing the culled chips that didn't make the KS cut.

 

[J7SC_Orion]

23 minutes ago, Mr. Fox said:

Too funny, bro. Great minds do think alike. I posted that over here earlier today.

 

I did not really expect the KS to be a different product as much as a binned sample of an existing product. So, there is some basis for disappointment. However, taking the glass is half full view, buying a KS might be the best way to avoid being dealt a tragic ticket in the silicon lottery. That's not a bad thing per se, but paying $200 extra for something that isn't as good or better than a superior 12900K silicon sample is where the basis for disappointment comes in. It's more like damage control  or loss prevention than an upgrade. You have to decide whether it is better to roll the dice on a K and hope for a win, or pay extra for something that is, at least, pretty good... with potential to be better than pretty good. If the chips are now being culled in favor of KS branding, then the chances of winning with a K might be drying up as old stock gets sold and replaced with stock representing the culled chips that didn't make the KS cut.

 

 

I think you're right about the culled 12900K, so KS is really a price hike in disuguise...at the end of the day, the 12900KS (and for that matter the AMD 5800X3D) are a blip, a sideshow until the main event later in 2022 (subject to various crisis not messing up the supply chains): The release of AMD's AM5 and Intel's Raptor Lake, then on to Meteor Lake (and whatever AMD offers in return).

 

By that time, more mature high-speed, low-latency DDR5 should also be floating about, and PCIe 5.0 GPUs and NVMEs will demand attention...

[schuck6566]

2 hours ago, J7SC_Orion said:

 

I think you're right about the culled 12900K, so KS is really a price hike in disuguise...at the end of the day, the 12900KS (and for that matter the AMD 5800X3D) are a blip, a sideshow until the main event later in 2022 (subject to various crisis not messing up the supply chains): The release of AMD's AM5 and Intel's Raptor Lake, then on to Meteor Lake (and whatever AMD offers in return).

 

By that time, more mature high-speed, low-latency DDR5 should also be floating about, and PCIe 5.0 GPUs and NVMEs will demand attention...

lol, I'm sure they can squeeze 1 more version out,like the 12900KSE...  (I'm the EXTREME SUPER K chip!)

[Storm-Chaser]

3 hours ago, schuck6566 said:

lol, I'm sure they can squeeze 1 more version out,like the 12900KSE...  (I'm the EXTREME SUPER K chip!)

[Storm-Chaser]

Probably wont be able to make anything happen but it's a bit colder this morning than it was the other day so I am going to give it a go. Report back later with results

[Mr. Fox]

5 hours ago, schuck6566 said:

lol, I'm sure they can squeeze 1 more version out,like the 12900KSE...  (I'm the EXTREME SUPER K chip!)

Each year replace  two e-cores with two p-cores and viola... 4 more generations. Planned obsolescence... er, I mean, Scalable architecture.

[Storm-Chaser]

10 minutes ago, Mr. Fox said:

Each year replace  two e-cores with two p-cores and viola... 4 more generations. Planned obsolescence... er, I mean, Scalable architecture.

It's looking like Intel has us stuck in a real world matrix like scenario. Oh wait, we already are, they are just perfecting their craft. They do the same thing with sockets to make us go out and get new motherboards when we want to upgrade the CPU.

 

Considering you can still get a single core CPUz benchmark score of 385 from an overclocked Q9650 @ 4.5GHz, I'm not impressed with their bad acting. 

 

[Mr. Fox]

2 hours ago, Storm-Chaser said:

It's looking like Intel has us stuck in a real world matrix like scenario. Oh wait, we already are, they are just perfecting their craft. They do the same thing with sockets to make us go out and get new motherboards when we want to upgrade the CPU.

 

Considering you can still get a single core CPUz benchmark score of 385 from an overclocked Q9650 @ 4.5GHz, I'm not impressed with their bad acting. 

 

I suppose it is all relative. I am not impressed by anything from Intel, AMD or NVIDIA. They all suck, and I simply choose which one seems to suck less. We can toss Micro$lop's name into the hat as well. They're not really PC hardware experts, which is evident by their Surface crap, but they excel at suckiness in the realm of technology.

[Storm-Chaser]

8 minutes ago, Mr. Fox said:

I suppose it is all relative. I am not impressed by anything from Intel, AMD or NVIDIA. They all suck, and I simply choose which one seems to suck less. We can toss Micro$lop's name into the hat as well. They're not really PC hardware experts, which is evident by their Surface crap, but they excel at suckiness in the realm of technology.

Maybe so but it still doesn't diminish the thrill of overclocking, at least in my opinion. I know with todays state of the art processors, overclocking is almost a lost, dark art because there is plenty of power right there in stock form. Regardless of their media sensationalism I still have plenty of respect for AMD and Intel.  Competition drives us all, right?

 

EDIT: And let me be clear, I hold nothing against either company. They are competent and every commercial enterprise including IT do the same things. 

 

So I was wondering about your chiller. I wanted to get details on that run. Are you ambient cooling or do you have an actual chiller like proper.

Edited April 6, 2022 by Storm-Chaser

[Mr. Fox]

36 minutes ago, Storm-Chaser said:

Maybe so but it still doesn't diminish the thrill of overclocking, at least in my opinion. I know with todays state of the art processors, overclocking is almost a lost, dark art because there is plenty of power right there in stock form. Regardless of their media sensationalism I still have plenty of respect for AMD and Intel.  Competition drives us all, right?

 

EDIT: And let me be clear, I hold nothing against either company. They are competent and every commercial enterprise including IT do the same things. 

 

So I was wondering about your chiller. I wanted to get details on that run. Are you ambient cooling or do you have an actual chiller like proper.

Indeed. Overclocking and benching is perhaps the only aspect of computing that keeps me interested in technology. Absent that, I don't think I would actually care about any of it anymore.

 

My setup is sort of a hybrid system. I have the HC-500A Chiller connected in series. Most of the time it is not used, and water simply circulates through it. It adds capacity to the loop, and that alone is beneficial for ordinary use. It takes longer to heat up the coolant because there is roughly twice the volume of the loop minus the chiller. When the ambient temperatures are hotter than ideal, I sometimes run the chiller to improve thermals for normal daily use. Living in the Phoenix area, summer heat can get my office ridiculously hot even with the central AC going with two overclocked PCs belching hot air into the room for days at a time. I don't game much anymore, but it was particularly useful for the occasional 12-16 hour gaming binge that kept the system under load for extended periods.

My radiator on that system is a MO-RA 360. The lines are connected with QDC fittings. When I push the overclocking harder for benching, I bypass the radiator completely because it is an impediment. It actually prevents the water from getting as cold as it needs to be because the radiator is designed to normalize the water temperature with ambient temperatures.

 

I have QDC fittings on the inlet port (female) and outlet port (male) of the radiator. I unplug the inlet fitting coming into the radiator, unplug the outlet fitting coming out of the radiator, then connect the ends of those lines directly to one another when I am on a benching spree. The QDC fittings are installed so that the orientation matches at the end of the lines where they connect to the radiator. So, the conversion from radiator and chiller to chiller-only takes about 15-20 seconds. I have a D5 pump mounted to the chiller and two D5 pumps with 270mm reservoirs mounted on the sides of the radiator. All three D5 pumps are working regardless of whether or not water is being circulated through the radiator.

 

The arrangement of the loop is like this:

 

Daily Use: CPU > GPU > Radiator > D5a > Chiller > D5b > D5c > CPU

Benching: CPU > GPU > D5a > Chiller > D5b > D5c > CPU

 

 

Edited April 6, 2022 by Mr. Fox

[Sir Beregond]

1 hour ago, Mr. Fox said:

I suppose it is all relative. I am not impressed by anything from Intel, AMD or NVIDIA. They all suck, and I simply choose which one seems to suck less. We can toss Micro$lop's name into the hat as well. They're not really PC hardware experts, which is evident by their Surface crap, but they excel at suckiness in the realm of technology.

100% the boat I am in right now.

 

Intel stagnated forever which gave room for AMD to catch up, but once AMD had it going, they were just as easily ready to pull the same sort of anti-consumer crap. Like why didn't we have non-X Zen 3 parts all the way up until now?

 

As for Nvidia, I am not impressed by rumored 500-600W cards. That's atrocious and if you need that much power to beat out your last gen and the competition, you are doing something wrong. If we just take performance relative to power draw, that sounds like a really crappy gen to gen uplift to me.

[Storm-Chaser]

26 minutes ago, Mr. Fox said:

Indeed. Overclocking and benching is perhaps the only aspect of computing that keeps me interested in technology. Absent that, I don't think I would actually care about any of it anymore.

 

My setup is sort of a hybrid system. I have the HC-500A Chiller connected in series. Most of the time it is not used, and water simply circulates through it. It adds capacity to the loop, and that alone is beneficial for ordinary use. It takes longer to heat up the coolant because there is roughly twice the volume of the loop minus the chiller. When the ambient temperatures are hotter than ideal, I sometimes run the chiller to improve thermals for normal daily use. Living in the Phoenix area, summer heat can get my office ridiculously hot even with the central AC going with two overclocked PCs belching hot air into the room for days at a time. I don't game much anymore, but it was particularly useful for the occasional 12-16 hour gaming binge that kept the system under load for extended periods.

My radiator on that system is a MO-RA 360. The lines are connected with QDC fittings. When I push the overclocking harder for benching, I bypass the radiator completely because it is an impediment. It actually prevents the water from getting as cold as it needs to be because the radiator is designed to normalize the water temperature with ambient temperatures.

 

I have QDC fittings on the inlet port (female) and outlet port (male) of the radiator. I unplug the inlet fitting coming into the radiator, unplug the outlet fitting coming out of the radiator, then connect the ends of those lines directly to one another when I am on a benching spree. The QDC fittings are installed so that the orientation matches at the end of the lines where they connect to the radiator. So, the conversion from radiator and chiller to chiller-only takes about 15-20 seconds. I have a D5 pump mounted to the chiller and two D5 pumps with 270mm reservoirs mounted on the sides of the radiator. All three D5 pumps are working regardless of whether or not water is being circulated through the radiator.

 

The arrangement of the loop is like this:

 

Daily Use: CPU > GPU > Radiator > D5a > Chiller > D5b > D5c > CPU

Benching: CPU > GPU > D5a > Chiller > D5b > D5c > CPU

 

 

Thanks for the comprehensive write up. Have any pictures on hand?

 

Supposing I want to cool with dry ice? Not sure if you are familiar with my ghetto cooler but it's pretty effective (because I live in upstate NY I can capitalize on freezing conditions. I have a plate heat exchanger that sits in a super cooled 5 gallon bucket of methanol. This way, I cool the entire heat exchanger, not just the other circuit (cold side). Amazingly enough, I dont even pump water through the cold side, it just gets immersed completely and perhaps not doing the circulation thing is a performance penalty but that just means down the road I can fit it on and perhaps pick up a few more MHz.

 

The plate heat exchanger was chosen because:

a) it's superior to air, liquid to liquid 

b) it has exponentially more surface area than a radiator. 

 

So this heat exchanger is then graphited into my conventional loop right before the intake on the block. 

 

Back to the question on hand, can I utilize what I have already in place to create a good environment for that type of chilling? i.e. surround the plate heat exchanger with dry ice in the 5 gallon bucket? 

 

My Setup:

 

 

LOL this was a run on a -14*F night. took about two hours to remove the heat exchanger because it was in a complete chunk of ice. Needless to say, it was performing very well that night.

 

 

 

 

 

 

[Mr. Fox]

3 minutes ago, Sir Beregond said:

100% the boat I am in right now.

 

Intel stagnated forever which gave room for AMD to catch up, but once AMD had it going, they were just as easily ready to pull the same sort of anti-consumer crap. Like why didn't we have non-X Zen 3 parts all the way up until now?

 

As for Nvidia, I am not impressed by rumored 500-600W cards. That's atrocious and if you need that much power to beat out your last gen and the competition, you are doing something wrong. If we just take performance relative to power draw, that sounds like a really crappy gen to gen uplift to me.

That is why overclocked benching is fun and interesting to me. I don't have to care about anything except winning and losing. Efficiency becomes an irrelevant topic. If the inefficient product wins, it's still a win. The more efficient product was slower. That is all that matters. Performing "almost" as good while using substantially less power is still losing. No trophies for participation and no honorable mention. Keep it simple: yes/no, on/off, up/down, win/lose, left/right, right/wrong... that is how I like it. 

 

That said, it depends on what one is measuring that defines the win or the loss. Doing more with less is not overclocked benching. It is something else, and the winner might be a different participant in that contest.

[Sir Beregond]

4 minutes ago, Mr. Fox said:

That is why overclocked benching is fun and interesting to me. I don't have to care about anything except winning and losing. Efficiency becomes an irrelevant topic. If the inefficient product wins, it's still a win. The more efficient product was slower. That is all that matters. Performing "almost" as good while using substantially less power is still losing. No trophies for participation and no honorable mention. Keep it simple: yes/no, on/off, up/down, win/lose, left/right, right/wrong... that is how I like it. 

 

That said, it depends on what one is measuring that defines the win or the loss. Doing more with less is not overclocked benching. It is something else, and the winner might be a different participant in that contest.

Sure, I get that. I am more speaking in general, not necessarily when doing OC sessions and benchmarking.

 

If I am gaming and a theoretical 4080 is drawing 500W, I think that's a loss of one sort for sure. 

 

But I realize a lot of enthusiasts likely don't agree. And that's fine.

Edited April 6, 2022 by Sir Beregond

[Mr. Fox]

35 minutes ago, Storm-Chaser said:

Thanks for the comprehensive write up. Have any pictures on hand?

 

Supposing I want to cool with dry ice? Not sure if you are familiar with my ghetto cooler but it's pretty effective (because I live in upstate NY I can capitalize on freezing conditions. I have a plate heat exchanger that sits in a super cooled 5 gallon bucket of methanol. This way, I cool the entire heat exchanger, not just the other circuit (cold side). Amazingly enough, I dont even pump water through the cold side, it just gets immersed completely and perhaps not doing the circulation thing is a performance penalty but that just means down the road I can fit it on and perhaps pick up a few more MHz.

 

The plate heat exchanger was chosen because:

a) it's superior to air, liquid to liquid 

b) it has exponentially more surface area than a radiator. 

 

So this heat exchanger is then graphited into my conventional loop right before the intake on the block. 

 

Back to the question on hand, can I utilize what I have already in place to create a good environment for that type of chilling? i.e. surround the plate heat exchanger with dry ice in the 5 gallon bucket? 

 

My Setup:

 

 

LOL this was a run on a -14*F night. took about two hours to remove the heat exchanger because it was in a complete chunk of ice. Needless to say, it was performing very well that night.

 

 

 

 

 

 

I will grab some pictures for you. I had to order more fittings after reconfiguring my setup with the distribution manifold and ran out of fittings in the process. As soon as I get the chiller reconnected I will post them. I should have the fittings I ordered by this weekend.

 

Conceptually, your setup works exactly the same except for your bucket with the submerged heat exchanger takes the place of the water chiller. It might be "ghetto" as you described it, but I am sure it works as well as, or maybe even better than, the chiller for a fraction of the cost. It might not be as convenient, but still seems highly effective. 

24 minutes ago, Sir Beregond said:

Sure, I get that. I am more speaking in general, not necessarily when doing OC sessions and benchmarking.

 

If I am gaming and a theoretical 4080 is drawing 500W, I think that's a loss of one sort for sure. 

 

But I realize a lot of enthusiasts likely don't agree. And that's fine.

Yes, I understand completely and we actually agree on that.

 

In some respects these are two separate topics. They are related in some ways, unrelated in other ways.

 

Some people are like snipers and only need one bullet for the head shot. Others go in spraying bullets everywhere, or they just use napalm or nukes and destroy everything in sight. God can sort them out later. Both are effective at eliminating opposition.

Edited April 6, 2022 by Mr. Fox

[J7SC_Orion]

...part of the problem is that modern chip design (CPU and even GPU) rely on branch prediction and longer pipelines to gain 'IPC'. Greatly oversimplified, but this is however also where a lot of vulnerabilities are injected...on the Ryzen 5k series, L3 cache speeds have now lost close to 40% just due to mitigations, similar if not worse story for past Intel. Even NVidia has had issues and fixes with / for that.

 

As a result, performance gains require an extra wallop of PL, unfortunately.

 

...while on a different layer, I wonder how long it will take before resizable_BAR and DirectStorage are found to have related weaknesses.

 

 

[Mr. Fox]

2 hours ago, J7SC_Orion said:

...part of the problem is that modern chip design (CPU and even GPU) rely on branch prediction and longer pipelines to gain 'IPC'. Greatly oversimplified, but this is however also where a lot of vulnerabilities are injected...on the Ryzen 5k series, L3 cache speeds have now lost close to 40% just due to mitigations, similar if not worse story for past Intel. Even NVidia has had issues and fixes with / for that.

 

As a result, performance gains require an extra wallop of PL, unfortunately.

 

...while on a different layer, I wonder how long it will take before resizable_BAR and DirectStorage are found to have related weaknesses.

All very true. I tend to not fret much over "security" issues and vulnerabilities because I am very cynical and not influenced much by fear and risk. So, whenever possible I disable mitigations.

 

IT Security is very similar to Big Pharma. Their ability to thrive depends upon always having a crisis and avoiding solutions. They make gobs of money off of the public perception of crisis and lose money the moment a cure surfaces. I truly believe most of the Boogie Man issues we face today are fabricated for the purpose of generating wealth for the self-anointed "leaders" of our world. They aren't really leaders, as much as legends in their own minds with nefarious intentions.

Edited April 6, 2022 by Mr. Fox

[J7SC_Orion]

6 minutes ago, Mr. Fox said:

All very true. I tend to not fret much over "security" issues and vulnerabilities because I am very cynical and not influenced much by fear and risk. So, whenever possible I disable mitigations.

 

IT Security is very similar to Big Pharma. Their ability to thrive depends upon always having a crisis and avoiding solutions. They make gobs of money off of the public perception of crisis and lose money the moment a cure surfaces. I truly believe most of the Boogie Man issues we face today are fabricated for the purpose of generating wealth for the self-anointed "leaders" of our world. They aren't really leaders, as much as legends in their own minds with nefarious intentions.

 

...apart from AMD driver updates etc, the problem is Microsoft Win 10 updates - some are necessary, others just drill deeper into your personal files...hard to tell which one is which when unblocking updates for truly important ones....KB123...321 doesn't really say very much... 

 

All that said, I keep a 'ghosted' drive from early of a completed and tested new setup handy, ie. one which isn't subsequently polluted with all this stuff. 

[Storm-Chaser]

33 minutes ago, Mr. Fox said:

All very true. I tend to not fret much over "security" issues and vulnerabilities because I am very cynical and not influenced much by fear and risk. So, whenever possible I disable mitigations.

 

IT Security is very similar to Big Pharma. Their ability to thrive depends upon always having a crisis and avoiding solutions. They make gobs of money off of the public perception of crisis and lose money the moment a cure surfaces. I truly believe most of the Boogie Man issues we face today are fabricated for the purpose of generating wealth for the self-anointed "leaders" of our world. They aren't really leaders, as much as legends in their own minds with nefarious intentions.

you just nailed it mr fox

[Mr. Fox]

9 hours ago, Storm-Chaser said:

The plate heat exchanger was chosen because:

a) it's superior to air, liquid to liquid 

b) it has exponentially more surface area than a radiator. 

It your "heat exchanger" the silver/aluminum thing with white plastic caps on four top ports? Do you have a link to where you purchased that so I can examine it closer?