*Official Benchmark Thread* - Post it here or it didn't happen :D

[Custom90gt]

Gone are the days when the customer mattered and companies stand behind their products. Intel would have been happy to sell you their overclocking warranty had they not ended it because of its lack of popularity:

https://www.intel.com/content/www/us/en/support/articles/000058556/processors.html

 

I remember when I worked as a tech while in undergrad and Intel would overnight me motherboards or CPUs that had gone bad in customer's systems (one reason we built our systems with MBs made by Intel).  No real questions asked, no ridiculous troubleshooting steps required.

[Mr. Fox]

2 hours ago, Custom90gt said:

Gone are the days when the customer mattered and companies stand behind their products. Intel would have been happy to sell you their overclocking warranty had they not ended it because of its lack of popularity:

https://www.intel.com/content/www/us/en/support/articles/000058556/processors.html

 

I remember when I worked as a tech while in undergrad and Intel would overnight me motherboards or CPUs that had gone bad in customer's systems (one reason we built our systems with MBs made by Intel).  No real questions asked, no ridiculous troubleshooting steps required.

Yes, I used to buy it. It was inexpensive. It was easy money for them because it was seldom needed as their processors were reliable and resilient. That reality was likely the reason it wasn't popular as most probably viewed it as being unnecessary. That is how insurance works. It is a waste of money until you need it, and then it becomes priceless. Needing it and not having it is tragic, especially when you're a wage-earner that isn't leeching an existence off of taxpayer sweat.

[Papusan]

20 hours ago, Papusan said:

From what I have seen... 13900K will do sub 15K in CBR20. But AMDs coming 7950X chips need to increase the performance with ++40% to reach 15K as well. Remember the 7700X needed 26% performance increase to reach the CBR20 result over older 5800X. 

 

See also... The need for +65% performance uplift to compete (or 48% performance uplift vs a castrated Raptor lake) https://www.notebookcheck.net/Intel-Core-i9-13900K-demolishes-the-Ryzen-9-5950X-with-an-up-to-67-lead-in-multi-core-Cinebench-test-but-there-is-a-huge-power-cost.639516.0.html

 

Not so sure 7700X will be a real 5800x successor, but from what I have seen, AMD won't push out 7800X at launch date. Not sure whats up. Maybe they want to go for the cheap Gaming king to grab attention first and later offer a real 5800 replacement at more premium price point. 

AMD Ryzen 9 7950X “Alleged” Zen 4 CPU Benchmarks Leak Out, 16 Core Up To 40% Faster Than 5950X

 

So in the best-case scenario, the AMD Ryzen 9 7950X will be slightly over 40% faster in multi-thread performance versus the Ryzen 9 5950X while the Ryzen 7 7700X is a little over 20% faster than the Ryzen 7 5800X.

 

Around 40% uplift in performance is slightly too little to make the 7950X becoming the multi core performance king (the price for the 16 core chips will rumored to be above 800$). So AMD need at least give it +46% or better say 50% performance increase over older 5950X to defend the price and stay on top in multicore performance. 

 

 

[electrosoft]

2 hours ago, Papusan said:

AMD Ryzen 9 7950X “Alleged” Zen 4 CPU Benchmarks Leak Out, 16 Core Up To 40% Faster Than 5950X

 

So in the best-case scenario, the AMD Ryzen 9 7950X will be slightly over 40% faster in multi-thread performance versus the Ryzen 9 5950X while the Ryzen 7 7700X is a little over 20% faster than the Ryzen 7 5800X.

 

Around 40% uplift in performance is slightly too little to make the 7950X becoming the multi core performance king (the price for the 16 core chips will rumored to be above 800$). So AMD need at least give it +46% or better say 50% performance increase over older 5950X to defend the price and stay on top in multicore performance. 

 

 

 

So single core (and overall architecture) goes to Intel again this round but the heat and power to get there is another hybrid mess so 16 / 32 AMD matches Intel's hybrid chip....

 

If Intel could get their node in order and offer a real, true 16 / 32 chip it would be no contest across all spectrum but it is what it is I guess.

 

Considering what a monster of a gaming chip the 5800X3D is, I have great expectations for the 7000 variants.

 

 

[electrosoft]

13 hours ago, Mr. Fox said:

For every action, there is an equal and opposite reaction. In contrast to Arctic Cooling's handling of this, we see examples of the opposite.

@johnkssshas always shared the belief that honesty is always the best policy. It seems that might not work for any of us going forward. We may have to stoop to their level to get the job done.  

 

Corporations are not your friend. Being honest will only get you shanked in the end as you just experienced. For a company that actively promotes and embraces overclocking to sell their products or get mind share and postings this really is a slap in the face. I can see XOC and obviously delidding but I tend to agree their chips aren't as robust as they used to be and the returns may have been hitting the bottom line a little harder than normal.

 

 

 

[Mr. Fox]

22 minutes ago, electrosoft said:

 

Corporations are not your friend. Being honest will only get you shanked in the end as you just experienced. For a company that actively promotes and embraces overclocking to sell their products or get mind share and postings this really is a slap in the face. I can see XOC and obviously delidding but I tend to agree their chips aren't as robust as they used to be and the returns may have been hitting the bottom line a little harder than normal.

Yeah, you are spot on about getting shanked. You cannot be honest and deal honorably with a party that is dishonest and has no honor. For the most part, the tech industry is dominated by companies and people of that low caliber of integrity. China being a key player in that industry probably has something to do with the problem as well.

 

Hard to say what the motivation is. It could also be that they noticed that AMD uses overclocking as a lame excuse to void a CPU warranty and thought they may as well take the low road like AMD and save a few bucks by using the same excuse. One thing for certain is nothing seems to be made with the level of quality that used to be normal. It does not matter what it is, (tech, cars, homes, appliances... everything) quality continues to decline as prices increase.

 

I may have to do a talking head feature video on my YouTube channel about this experience and share the email information with a broader audience. Maybe GamersNexus or WCCFTech will spot it and add some gasoline to the fire.

[Mr. Fox]

On 8/25/2022 at 10:45 AM, electrosoft said:

Happy belated Birthday @Papusan! It's wonderful when kids turn into adults of their own. They're still your kids but they also become your friend on a level too.

Are there any pictures of the @Papusangaming desk to share yet?

[Papusan]

26 minutes ago, Mr. Fox said:

Are there any pictures of the @Papusangaming desk to share yet?

Not yet. Need some help from my son. And my wife struggle with the health equal as for me right now. 

[Gumwars]

Howdy all, long time no see.

 

This is a benchmark thread, right?

 

How's this?

[Mr. Fox]

11 minutes ago, Gumwars said:

Howdy all, long time no see.

 

This is a benchmark thread, right?

 

How's this?

A lot better than this.

https://www.3dmark.com/pr/1694278

 

[Papusan]

On 8/25/2022 at 7:08 AM, Papusan said:

Regarding cooling. I will ask for the replacement parts for my AC 420 then put it in the other desktop build. I just ordred a few cooling parts for my next cooling setup on the newer machine. And yep, the added tax+the awful monetary value here home make it more expensive than it has to be. There is no pc-parts shops here home that have much cooling parts. Everything need to be ordered outside the country. I will buy some more parts on the used market here home. As well re-use some parts I already have.

Got a used EK XE480 rad on the used market here home. This is an EOL model from EK but this one should still be in the top regarding cooling performance. Got it for around half the price vs Ebay or a  new similar 60mm thick 480 rad from EK/Alphacool. Just sad I didn't buy both of them 1.5 week ago. Now I just need find a twin. So it will be 2x480 in the front and 1x420 at the top. God bless I bougth a big enough chassis, LOOL

 

I think big is out. People want smaller and crampy chassis nowadays. Some call they even sexy. Hence a bigger chance for all to find used bigger rads. 

 

 

Here's the successor to the XE 480. And I had to paint over the awful silver/steel color to match my all black build. All has to be black as you know 🙂

https://www.ekwb.com/shop/ek-quantum-surface-x480m-black

 

 

[electrosoft]

Asrock A380 w/ newest beta drivers puts it right in between a 1650 and 1650 Super at this point. Testing it on the "new realistic" system I'm setting up for my son in law once financial realities kicked in and he adjusted his expectations commiserate with his cash flow status. 😁

 

 

 

[johnksss]

22 hours ago, Mr. Fox said:

For every action, there is an equal and opposite reaction. In contrast to Arctic Cooling's handling of this, we see examples of the opposite.

@johnkssshas always shared the belief that honesty is always the best policy. It seems that might not work for any of us going forward. We may have to stoop to their level to get the job done.  

And for these reasons I try to stay with buying all my parts from MC until they change their policy on their warranty. 

[Mr. Fox]

14 minutes ago, johnksss said:

And for these reasons I try to stay with buying all my parts from MC until they change their policy on their warranty. 

Maybe someday I will be blessed with a MC within 100 miles of where I live. I hope so. The closest ones are in Tustin, CA (yours) and Dallas. I have suggested they look at adding one in Phoenix. I really don't know why there is not one here already. It is a gigantic metropolis. It would be wildly successful here, and Arizona is very business-friendly with unusually low taxes compared to many of the cities that are almost hostile toward businesses and tax-Nazis where they have locations.

 

I am still waiting for the guy from Intel to send me an official copy of the warranty that states the act of overclocking the processor or memory automatically voids the warranty for failures of any kind, whether caused by overclocking or not.

[johnksss]

17 minutes ago, Mr. Fox said:

Maybe someday I will be blessed with a MC within 100 miles of where I live. I hope so. The closest ones are in Tustin, CA (yours) and Dallas. I have suggested they look at adding one in Phoenix. I really don't know why there is not one here already. It is a gigantic metropolis. It would be wildly successful here, and Arizona is very business-friendly with unusually low taxes compared to many of the cities that are almost hostile toward businesses and tax-Nazis where they have locations.

At the current time, they seem to have no plans on expanding..... Unfortunately.

 

Side note:

I did get stuck with a 980X for 3 years because it had a burnt pad and at the time Microcenter nor Intel would not replace it. Intel came back with their tuning plan and I got it replaced at that point after some back and forth about it.

[Mr. Fox]

12 hours ago, johnksss said:

At the current time, they seem to have no plans on expanding..... Unfortunately.

 

Side note:

I did get stuck with a 980X for 3 years because it had a burnt pad and at the time Microcenter nor Intel would not replace it. Intel came back with their tuning plan and I got it replaced at that point after some back and forth about it.

I had to just bite the bullet and order another CPU. It should be here on Tuesday. The failing 12900KS seems to work right most of the time other than the dead memory channel. Very odd. No visibly burned pads or anything else. I am guessing something internal in the transistor circuitry. I was hoping it would work fine on the DDR4 Strix mobo and I would just live with knowing it did not work right for DDR5, but no such luck. What I mean by most of the time is that it will randomly throw a BSOD. It did that three times today within the span of 30 minutes around mid-morning, but worked correctly otherwise from 6:30 AM when I started work until I stopped working at 4:00 PM. When it does the BSOD the mobo lights act weird and it behaves like it is retraining the memory, so it might be interesting to see if it keeps working long enough for the new 12900KS to arrive on Tuesday. If not, I may have to use my laptop for work for a few days.

 

Then I will have to decide whether to keep the Strix and sell the Unify-X or vice versa. Part of me wonders if the MSI board did something to the CPU like the Z490 Apex that killed my first 10900KF. I like the Unify-X way better than I do the Strix D4, but I wonder if it wouldn't be the safer choice to keep the Strix. The Unify-X has had occasional random weirdness all along, but it is an excellent motherboard when it is not doing something goofy. First the memory SPD corruption, now the CPU losing a memory channel. Hard to feel like continuing to use it is the smart thing to do at this point. And, it is my second Unify-X, so that makes it even harder to have confidence in it being a reliable product.

[jaybee83]

22 hours ago, electrosoft said:

 

So single core (and overall architecture) goes to Intel again this round but the heat and power to get there is another hybrid mess so 16 / 32 AMD matches Intel's hybrid chip....

 

If Intel could get their node in order and offer a real, true 16 / 32 chip it would be no contest across all spectrum but it is what it is I guess.

 

Considering what a monster of a gaming chip the 5800X3D is, I have great expectations for the 7000 variants.

 

 

 

indeed. there were some leaked internal preliminary tests by AMD that showed the following:

1st gen 3D-VCache 5800X3D vs. 5800X = +15% faster

2nd gen 3D-VCache 7000X3D vs. 7000X = +30% faster

 

imagine Intel's reaction when AMD brings 30% faster gaming cpus just a few short months after Raptor Lake hits the shelves, which in turn are only a few % faster than non-3D Zen 4 cpus 😄 

 

its definitely an awesome time for PC enthusiasts with lots of competition going on.

[johnksss]

13 hours ago, Mr. Fox said:

I had to just bite the bullet and order another CPU. It should be here on Tuesday. The failing 12900KS seems to work right most of the time other than the dead memory channel. Very odd. No visibly burned pads or anything else. I am guessing something internal in the transistor circuitry. I was hoping it would work fine on the DDR4 Strix mobo and I would just live with knowing it did not work right for DDR5, but no such luck. What I mean by most of the time is that it will randomly throw a BSOD. It did that three times today within the span of 30 minutes around mid-morning, but worked correctly otherwise from 6:30 AM when I started work until I stopped working at 4:00 PM. When it does the BSOD the mobo lights act weird and it behaves like it is retraining the memory, so it might be interesting to see if it keeps working long enough for the new 12900KS to arrive on Tuesday. If not, I may have to use my laptop for work for a few days.

 

Then I will have to decide whether to keep the Strix and sell the Unify-X or vice versa. Part of me wonders if the MSI board did something to the CPU like the Z490 Apex that killed my first 10900KF. I like the Unify-X way better than I do the Strix D4, but I wonder if it wouldn't be the safer choice to keep the Strix. The Unify-X has had occasional random weirdness all along, but it is an excellent motherboard when it is not doing something goofy. First the memory SPD corruption, now the CPU losing a memory channel. Hard to feel like continuing to use it is the smart thing to do at this point. And, it is my second Unify-X, so that makes it even harder to have confidence in it being a reliable product.

All I can say to the "bite the bullet" is I have done than quite a few times over the years. Which resulted in thousands and thousands of dollars lost in the time span from 2008 to 2022. One can not win them all and you take the wins you should not have gotten with the wins that you did and most of the time you are still far on top. This is pretty much true for most people....

 

As to the 12900KS, more than likely it is bad. As was my first 12900K when I basically had to replace everything. Motherboard/GPU/Power Supply/Memory and finally the CPU because it kept throwing errors after the new parts were installed.

 

Not really sure I would be using that for work as memory type errors are much harder to recover from if the system crashes one to many times. Playing games and encoding videos is one thing, but mission critical is quite another to in trust in a system that is not fairly stable with memory. And using a laptop is really not a down fall as it can be connected to you current keyboard, mouse and monitor. Convenient? Not really. Doable for a few days? Definitely.

 

And it is very well possible the MSI mother board did in fact kill the CPU, but that usually only happens when the pins touch or the socket had moisture in or around it and condensation or a drop touched the board in the wrong place. Or liquid was under neath any one of the heatsinks or pads that have to be taken completely apart to even figure that out. Boards can last fairly long with 50/50 but direct liquid is usually instant fail though. Case in point. When my system was sitting under and inch of 50/50 it was still running, but because I did not fully clean the board I lost everything on the second boot attempt. It wasn't until I stripped the board and used my microscope to see between the components that stuff was stuck between the legs. Along with all the other things wrong. Now I'm like 50-1 (They did not replace a Kingpin 1080 TI because they said it had water damage - So I had to pony up and buy another one.) with EVGA, but I did ask that since it was their water block that broke and cause this, would they cover the replacement parts. The answer they came back was yes if that was in fact true. After they tested the block, they came to the understanding it was. Once I found out that everyone else was going to replace the parts, I let EVGA know that it would not be needed. Only the GPU/Block and PSU, which they did..... But when something like this happen quite a few years ago...it was quite a different story that cost a few thousand dollars because of my mistake....

 

For the record I'm still on the same Unify-X. All MSI did was to repair it. Which I have no issue with as long as it works as intended, but when it comes back with broke pieces.....Then I have to use other means to make it right.

[Tenoroon]

@Mr. FoxI plan on sanding the heatpipes today or tommorow, and I need a bit of input on how and what to do...

I have some pretty coarse sandpaper (not very good with sandpaper, I don't what grit.) I sanded a bit of a test heatpipe to see how easy it is to remove the previous solder from the copper. I believe I need some finer sandpaper as it seems to sand the copper too, turning the area into a pinkish hue and makes the heatpipe look thinner.

Now, do I need to completely remove the solder from the copper (where you don't see the light gray from the solder) or do I just need to flatten the solder out (keep the light gray, but it looks smooth like in the pic below.)

If I need to completely remove any remnants of the solder, what kind of sandpaper would you recommend? Would you also recommend something different for sanding the black paint off of the fins?

[Mr. Fox]

9 hours ago, johnksss said:

All I can say to the "bite the bullet" is I have done than quite a few times over the years. Which resulted in thousands and thousands of dollars lost in the time span from 2008 to 2022. One can not win them all and you take the wins you should not have gotten with the wins that you did and most of the time you are still far on top. This is pretty much true for most people....

 

As to the 12900KS, more than likely it is bad. As was my first 12900K when I basically had to replace everything. Motherboard/GPU/Power Supply/Memory and finally the CPU because it kept throwing errors after the new parts were installed.

 

Not really sure I would be using that for work as memory type errors are much harder to recover from if the system crashes one to many times. Playing games and encoding videos is one thing, but mission critical is quite another to in trust in a system that is not fairly stable with memory. And using a laptop is really not a down fall as it can be connected to you current keyboard, mouse and monitor. Convenient? Not really. Doable for a few days? Definitely.

 

And it is very well possible the MSI mother board did in fact kill the CPU, but that usually only happens when the pins touch or the socket had moisture in or around it and condensation or a drop touched the board in the wrong place. Or liquid was under neath any one of the heatsinks or pads that have to be taken completely apart to even figure that out. Boards can last fairly long with 50/50 but direct liquid is usually instant fail though. Case in point. When my system was sitting under and inch of 50/50 it was still running, but because I did not fully clean the board I lost everything on the second boot attempt. It wasn't until I stripped the board and used my microscope to see between the components that stuff was stuck between the legs. Along with all the other things wrong. Now I'm like 50-1 (They did not replace a Kingpin 1080 TI because they said it had water damage - So I had to pony up and buy another one.) with EVGA, but I did ask that since it was their water block that broke and cause this, would they cover the replacement parts. The answer they came back was yes if that was in fact true. After they tested the block, they came to the understanding it was. Once I found out that everyone else was going to replace the parts, I let EVGA know that it would not be needed. Only the GPU/Block and PSU, which they did..... But when something like this happen quite a few years ago...it was quite a different story that cost a few thousand dollars because of my mistake....

 

For the record I'm still on the same Unify-X. All MSI did was to repair it. Which I have no issue with as long as it works as intended, but when it comes back with broke pieces.....Then I have to use other means to make it right.

I am not sure what I want to do as far as which mobo to keep. I really like the Unify-X, but not knowing if it is what damaged the memory and CPU, or if it was the CPU that damaged the memory makes it seem like maybe it shouldn't be worth the risk. The latter seems far less likely to me. I almost feel like whatever it did to screw up the memory also damaged the CPU and it just took longer for the evidence to show on the CPU. The Strix isn't a bad mobo by any means, but not even close to the same caliber as the Unify-X. The Strix is nice and the Unify-X is excellent, but I'd rather have a nice mobo than another round of parts killing.

 

Nothing is mission-critical on the work system in terms of risk for data loss. Everything I do, including the data part, is cloud-based. I could do everything with a Chromebook or tablet if I had to, even though I wouldn't really want to.

9 hours ago, Tenoroon said:

@Mr. FoxI plan on sanding the heatpipes today or tommorow, and I need a bit of input on how and what to do...

I have some pretty coarse sandpaper (not very good with sandpaper, I don't what grit.) I sanded a bit of a test heatpipe to see how easy it is to remove the previous solder from the copper. I believe I need some finer sandpaper as it seems to sand the copper too, turning the area into a pinkish hue and makes the heatpipe look thinner.

Now, do I need to completely remove the solder from the copper (where you don't see the light gray from the solder) or do I just need to flatten the solder out (keep the light gray, but it looks smooth like in the pic below.)

If I need to completely remove any remnants of the solder, what kind of sandpaper would you recommend? Would you also recommend something different for sanding the black paint off of the fins?

No you do not need to totally remove all of the solder, just level it out where it makes contact. The copper the heat pipes is made of is very soft and flexible. It's easy to sand it down and make it sound so you do want to avoid doing that. I would use something like 400 or 600 grit paper to take down the solder quickly and then move to 800 or 1000 grit to smooth it up and just don't sand on the bare copper that much. The only reason you need to sand that surface is so that there's no clumps or chunks of old solder or paint to interfere with the resoldering.

[Tenoroon]

26 minutes ago, Mr. Fox said:

No you do not need to totally remove all of the solder, just level it out where it makes contact. The copper the heat pipes is made of is very soft and flexible. It's easy to sand it down and make it sound so you do want to avoid doing that. I would use something like 400 or 600 grit paper to take down the solder quickly and then move to 800 or 1000 grit to smooth it up and just don't sand on the bare copper that much. The only reason you need to sand that surface is so that there's no clumps or chunks of old solder or paint to interfere with the resoldering.

Yeah, that's what I thought, but wasn't entirely sure.

I was digging through my dremel kit and found a "metal brush" that works really well. It gets the old solder off, and doesn't mess with the copper too much. Here's what I'm talking about:



The pipe I used in the pic was a "test" pipe from the GT75 heatsink that is already a bit bent. I didn't want to FUBAR a nice pipe that I may use in the future. Here's what one of the ends looks like after using the "metal brush" for a few minutes (I am aware of the solder clumps on the left):



 

[Reciever]

Well I have to hand it to evga, solid product from whomever they chose as their supplier. I think I figured out what was wrong. My server 1200w psu was not initialized so no power was being fed to the GPU's I hooked up, which is why it shorted. Like an idiot I tried to power it on again after the warning and that's when it gave up the ghost. 

 

What I'm really happy about though as compared to maybe 10 years ago, is it didn't take anything with it on its way out, or at least nothing I have tested so far. 

 

I replaced it with an evga 1300 g2, I was going to get the p2 but I read some issues of coil whine and the g2 could get here by today so it kind of just worked out that way. 

 

Also @Mr. Fox, thanks to your suggestion of the frame I don't think I can go back to standard cases anymore.

 

 

[johnksss]

[jaybee83]

10 hours ago, johnksss said:

interesting, so u havent beaten ur best 5950x scores yet in CB20? just judging by the saved scores i mean.

[johnksss]

20 minutes ago, jaybee83 said:

interesting, so u havent beaten ur best 5950x scores yet in CB20? just judging by the saved scores i mean.

Didn't know that I was suppose to?