Jump to content

@LinusTech You know what I would love, is a rig performance calculator published by LTT. I will even pay for the service of such a tool.

 

To give you an idea of what I am on about: A website with drop downs where i can select cpu, gpu, ram, ssd etc. Sliders to indicate the overclock that i wish to run(in regards with GPU, CPU and RAM) and your calculator spits out the performance you guys got testing this rig, You will probably need to get a full time employee for this, hence I am more than happy to pay for a service like this. Don't get me wrong, there are many sites that publish performance, but this tool should be so much more. If I want to select a Intel i3 with a titan X , I should be able to (extreme case for example). I want to select corsair vs kingston. I want to select gigabyte vs msi. It should be the biggest single database of parts, all tested together and with some components (GPU, CPU and RAM) overclocked

 

This is no small feat. The amount of iterations are unimaginable. But I don't care. Its time someone does this, and it might as well be LTT.

 

PS. Please have this done when the first Pascal's drop :-)

 

Cheers

Carel 

Link to post
Share on other sites

No chance m8. It would be less pricy if you go ahead and test the stuff urself. (After buying it ofcourse)

Budget Rig "Curable":     | FX 6300 @4.5Ghz | Asus R9 270x | Asus Crosshair IV Extreme | 16GB HyperX Beast | 120GB PNY SSD
 

Tablet "Buddy":                 Trekstor Wintron 10.1|

 

Link to post
Share on other sites

Corsair vs Kingston... so Hynix MFR vs Hynix MFR. Gotcha.

Link to post
Share on other sites

7 minutes ago, Carel101 said:

 

Psst... he's @LinusTech , @Linus is an impostor :P

 

This would take so long, you'd need a whole team of people and even then it'd take literally months, possibly even years. Plus, their work would never be done, since new stuff would always be coming out. Plus, there are a metric butt-ton of different types of benchmarks they'd have to run. It's just not going to happen, there are millions upon millions of hardware combinations and many many benchmarks to run on each. Would be useful though.

Project White Lightning (My ITX Gaming PC): Core i5-4690K | CRYORIG H5 Ultimate | ASUS Maximus VII Impact | HyperX Savage 2x8GB DDR3 | Samsung 850 EVO 250GB | WD Black 1TB | Sapphire RX 480 8GB NITRO+ OC | Phanteks Enthoo EVOLV ITX | Corsair AX760 | LG 29UM67 | CM Storm Quickfire Ultimate | Logitech G502 Proteus Spectrum | HyperX Cloud II | Logitech Z333

Benchmark Results: 3DMark Firestrike: 10,528 | SteamVR VR Ready (avg. quality 7.1) | VRMark 7,004 (VR Ready)

 

Other systems I've built:

Core i3-6100 | CM Hyper 212 EVO | MSI H110M ECO | Corsair Vengeance LPX 1x8GB DDR4  | ADATA SP550 120GB | Seagate 500GB | EVGA ACX 2.0 GTX 1050 Ti | Fractal Design Core 1500 | Corsair CX450M

Core i5-4590 | Intel Stock Cooler | Gigabyte GA-H97N-WIFI | HyperX Savage 2x4GB DDR3 | Seagate 500GB | Intel Integrated HD Graphics | Fractal Design Arc Mini R2 | be quiet! Pure Power L8 350W

 

I am not a professional. I am not an expert. I am just a smartass. Don't try and blame me if you break something when acting upon my advice.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

...why are you still reading this?

Link to post
Share on other sites

1 minute ago, harrynowl said:

Corsair vs Kingston... so Hynix MFR vs Hynix MFR. Gotcha.

lol

 

When it comes to RAM and motherboards, they've already proven that you're not gonna get higher FPS with a better motherboard. RAM might give you 1-2 FPS more, but that's about it. I think it should just be for pairing CPU to GPU, and amount of RAM.


But, it would take an insanely long time to test all the combos out there, especially going from GPU brand to GPU brand. Maybe they'd just test the GPU, and give you a rough estimate of performance for each cooler brand.

I used to be quite active here.

Link to post
Share on other sites

As much as I and many others would love a tool like this is nearly impossible to do such a thing as you'd have to benchmark every combination of GPU / CPU / RAM at various OC's and that is a near impossible task for any one person or team to handle, especially because different chips OC differently and then there are driver revisions etc.

Sergeant, United States Marine Corps

Network Administrator, Comptia A+, Security+, Cisco Certified Networking Associate

From a G3258 to dual Xeon E5-2670's

Link to post
Share on other sites

These are the dreams of a mad man!!!!

Cool, yes

but even after the size and time that you mentioned, not every i7-6600k or Ripjaw V 1600 4x4gb will perform the same as the other let alone considering it would be a few hundred or thousand combos for just cpu and mobo

The lga 2011 suits dozens of cpus multiply by the same number of boards then the dozens of types of ram and the quantities and then the hundreds of storage options, followed by the hundreds of gpu and psu options, you end up with a million combos for a single socket. 

https://en.wikipedia.org/wiki/LGA_2011

                     .
                   _/ V\
                  / /  /
                <<    |
                ,/    ]
              ,/      ]
            ,/        |
           /    \  \ /
          /      | | |
    ______|   __/_/| |
   /_______\______}\__}  

Spoiler

[i7-7700k@5Ghz | MSI Z270 M7 | 16GB 3000 GEIL EVOX | STRIX ROG 1060 OC 6G | EVGA G2 650W | ROSEWILL B2 SPIRIT | SANDISK 256GB M2 | 4x 1TB Seagate Barracudas RAID 10 ]

[i3-4360 | mini-itx potato | 4gb DDR3-1600 | 8tb wd red | 250gb seagate| Debian 9 ]

[Dell Inspiron 15 5567] 

 

 

Link to post
Share on other sites

7 minutes ago, Carel101 said:

 

PS. Please have this done when the first Pascal's drop :-)

You do realize something like this would take more than 2 months... right....?

|PSU Tier List /80 Plus Efficiency| PSU stuff if you need it. 

My system: PCPartPicker || For Corsair support tag @Corsair Josephor @Corsair Nick || My 5MT Legacy GT Wagon ||

Link to post
Share on other sites

not possible.

Intel Xeon E5 1650 v3 @ 3.5GHz 6C:12T / CM212 Evo / Asus X99 Deluxe / 16GB (4x4GB) DDR4 3000 Trident-Z / Samsung 850 Pro 256GB / Intel 335 240GB / WD Red 2 & 3TB / Antec 850w / RTX 2070 / Win10 Pro x64

HP Envy X360 15: Intel Core i5 8250U @ 1.6GHz 4C:8T / 8GB DDR4 / Intel UHD620 + Nvidia GeForce MX150 4GB / Intel 120GB SSD / Win10 Pro x64

 

HP Envy x360 BP series Intel 8th gen

AMD ThreadRipper 2!

5820K & 6800K 3-way SLI mobo support list

 

Link to post
Share on other sites

3 minutes ago, Kobathor said:

lol

 

When it comes to RAM and motherboards, they've already proven that you're not gonna get higher FPS with a better motherboard. RAM might give you 1-2 FPS more, but that's about it. I think it should just be for pairing CPU to GPU, and amount of RAM.


But, it would take an insanely long time to test all the combos out there, especially going from GPU brand to GPU brand. Maybe they'd just test the GPU, and give you a rough estimate of performance for each cooler brand.

Must everything come back to gaming? 

 

but yes, I'd say with the new drops approaching by the time they got caught up with the zen and pascal the next line would be coming out

                     .
                   _/ V\
                  / /  /
                <<    |
                ,/    ]
              ,/      ]
            ,/        |
           /    \  \ /
          /      | | |
    ______|   __/_/| |
   /_______\______}\__}  

Spoiler

[i7-7700k@5Ghz | MSI Z270 M7 | 16GB 3000 GEIL EVOX | STRIX ROG 1060 OC 6G | EVGA G2 650W | ROSEWILL B2 SPIRIT | SANDISK 256GB M2 | 4x 1TB Seagate Barracudas RAID 10 ]

[i3-4360 | mini-itx potato | 4gb DDR3-1600 | 8tb wd red | 250gb seagate| Debian 9 ]

[Dell Inspiron 15 5567] 

 

 

Link to post
Share on other sites

True. Brand name testing is overkill. As for drivers. As baseline driver is selected and as far as possible this gets used for every test. Only when a new architecture is launced should drivers be updated.

 

The start line should also be for hardware that can be purchased now as new . So no GTX680 for example.

Link to post
Share on other sites

1 minute ago, RedWulf said:

Must everything come back to gaming? 

What else do you think people would look it up for? I'd say a lot of professionals (other than people who work for small companies running YouTube channels, IE LTT or TechSyndicate) don't even know how to build a computer, let alone what a GTX 980 or i7-5960x is. So, they have people put their computers together for them.

 

The only people I see using this tool would be gamers. Things like video editing have their own tools (you can find CPU benchmarks for every CPU out there, and how good they are at rendering, ect) so this tool wouldn't need to incorporate them.


That means, the people using it would want to see how many FPS they can get in whatever game with a given setup, and that's all they'd want to know.

I used to be quite active here.

Link to post
Share on other sites

6 minutes ago, Carel101 said:

True. Brand name testing is overkill. As for drivers. As baseline driver is selected and as far as possible this gets used for every test. Only when a new architecture is launced should drivers be updated.

 

The start line should also be for hardware that can be purchased now as new . So no GTX680 for example.

It would be a impossible task, no matter how easy your mind think it is. Core i3 with Titan X. Do you know how many Core i3 and Titan X are out there. CPU OC, GPU OC, system configuration, bios setup, screen resolution, game settings, driver version, etc.

Intel Xeon E5 1650 v3 @ 3.5GHz 6C:12T / CM212 Evo / Asus X99 Deluxe / 16GB (4x4GB) DDR4 3000 Trident-Z / Samsung 850 Pro 256GB / Intel 335 240GB / WD Red 2 & 3TB / Antec 850w / RTX 2070 / Win10 Pro x64

HP Envy X360 15: Intel Core i5 8250U @ 1.6GHz 4C:8T / 8GB DDR4 / Intel UHD620 + Nvidia GeForce MX150 4GB / Intel 120GB SSD / Win10 Pro x64

 

HP Envy x360 BP series Intel 8th gen

AMD ThreadRipper 2!

5820K & 6800K 3-way SLI mobo support list

 

Link to post
Share on other sites

15 minutes ago, ThinkWithPortals said:

Psst... he's @LinusTech , @Linus is an impostor :P

 

This would take so long, you'd need a whole team of people and even then it'd take literally months, possibly even years. Plus, their work would never be done, since new stuff would always be coming out. Plus, there are a metric butt-ton of different types of benchmarks they'd have to run. It's just not going to happen, there are millions upon millions of hardware combinations and many many benchmarks to run on each. Would be useful though.

Psst... he doesn't have notifications turned on so tagging him does nothing :P

Solve your own audio issues  |  First Steps with RPi 3  |  Humidity & Condensation  |  Sleep & Hibernation  |  Overclocking RAM  |  Making Backups  |  Displays  |  4K / 8K / 16K / etc.  |  Do I need 80+ Platinum?

If you can read this you're using the wrong theme.  You can change it at the bottom.

Link to post
Share on other sites

Impossible no. Massive task, oh yes.
If you truly want to do this you might need some community input.

What bugs me, is that so many performance benchmark software has the ability to upload results, but they never have a combined searchable database that is baselined and error corrected

 

Take for example passmark. You can upload you results. But if you search their DB, you get isolated CPU performance or isolated GPU. That is just not flexible enough. I want to see the performance drop between 970 SLI on i5 vs i7. I want to see how far I should overclock the i5 to be on par with the i7 again. Now that is useful info to have at your fingertips when you are build a new pc.

Link to post
Share on other sites

1 hour ago, Carel101 said:

Impossible no. Massive task, oh yes.
If you truly want to do this you might need some community input.

What bugs me, is that so many performance benchmark software has the ability to upload results, but they never have a combined searchable database that is baselined and error corrected

 

Take for example passmark. You can upload you results. But if you search their DB, you get isolated CPU performance or isolated GPU. That is just not flexible enough. I want to see the performance drop between 970 SLI on i5 vs i7. I want to see how far I should overclock the i5 to be on par with the i7 again. Now that is useful info to have at your fingertips when you are build a new pc.

Take a look on HWBot.

Link to post
Share on other sites

HWBOT is a bit on the extreme overclock side, I really don't see the benefit of comparing liquid nitrogen cooled cards...  I am sure there is useful info on hwbot if you search long and hard. Nothing jumps out to quickly compare rigs.

Link to post
Share on other sites

1 hour ago, Carel101 said:

HWBOT is a bit on the extreme overclock side, I really don't see the benefit of comparing liquid nitrogen cooled cards...  I am sure there is useful info on hwbot if you search long and hard. Nothing jumps out to quickly compare rigs.

There's plenty of non-LN2 cooled cards there. But it's the best database for comparing a wide variety of configurations with detailed results (as you need to have a valid screenshot)

Link to post
Share on other sites

That's a good point. The data is available out there, for example at hwbot like to mention @harrynowl We just need a way to present it in a more logical format in the form of a calculator. So the real job is not for a performance tester but rather for a good BI / DBA specialist that can collate all the date and present it.

 

If we keep this up we will have the job spec nailed in no time, ;-)

Link to post
Share on other sites

Passmark is closest what you can get now. It has GPUs, SSDs, RAM and ofc CPUs. I know some people don't really think their benchmark does it well for gaming performance seekers, buts thats good baseline test.

 

Overclocking side is bit problematic. Silicon lottery is huge thing. There's always safe numbers per component. Like what everyone should be able to get with slight&safe voltage increase. But slider to show how high you can go... Not really seeing it. With GPUs, brand has some impact too.

 

In all, we are talking about more than one person doing full hours more than few months. And would have to continue doing so for as long as there is PC components being sold. Not really something that in scope of company whose main job is making digital media content (as in videos).

^^^^ That's my post ^^^^
<-- This is me --- That's your scrollbar -->
vvvv Who's there? vvvv

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×