Cinebench font (system info on left); older CPU benchmark requests (newest=Athlon 64 X2 4000+); i7-6700K says 3.5 GHz base (but runs @ 4); Handbrake…

PianoPlayer88Key · March 2, 2021

Hey all, was wondering about a few things mostly related to CPUs....

Does anyone know what font Cinebench (R23, although other versions may be okay too) uses in the section where it mentions your system specs? (I'm working on a screenshot collage for something - basically comparing the uplift of a hypothetical future CPU over an existing CPU, vs the existing CPU's uplift over an older CPU. Alternately maybe I could search for other screenshots that include the characters in the name that I need that my own system doesn't have, and copy/ accordingly.)

Does anyone by any chance have:

AMD Athlon 64 X2 4000+ Brisbane Dual-Core 2.1 GHz Socket AM2 65W ADO4000DDBOX Processor (with the stock CPU cooler and TIM)
GIGABYTE GA-MA69G-S3H AM2 AMD 690G HDMI ATX AMD Motherboard (includes on-board ATI Radeon Xpress 1250 graphics)
2x G.SKILL 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory Model F2-6400CL5D-2GBNQ

(Or at least the same CPU, graphics and DDR2-800 RAM even if it's a different motherboard and not the exact same RAM SKUs?)

That's what I used to have, but sent them to a recycler about 8 or so years ago after the motherboard had passed away. Therefore I'm obviously unable to benchmark them, but was wondering if someone else hopefully has them and could run a few benchmarks for me? :

Cinebench - R23, R20, R15, optionally R11.5, 10, 9.5, 2003, were there any before that?
3DMark - Ice Storm, and any older versions of 3DMark back to '99; were there any older ones as well?
7-Zip and/or other apps with benchmark capabilities
any other benchmarks you can think of, that can also be run on more modern systems, like
- i7-6700K, GTX 970M 6GB, 64GB DDR4-2133
- i7-4790K, GTX 1060 3GB / HD 4600 (one or the other), 32GB DDR3-1600

I've looked on HWBot and maybe a couple other places, but it looks like a lot of those may have been overclocked, or at least the scores are all over the place.

I'd really like the benchmarks to be done at completely factory-stock settings, basically the way I ran my system when I had it. (I don't even remember ever enabling any type of memory overclock / enhancements, like whatever AMD's version of XMP would have been at the time, or if anything like that was even available. Replacing failed TIM would be permissible, though, but preferably with something similar quality to the original; nothing fancy like NT-H1, Kryonaut, etc.)

Also does anyone have the following older hardware, who could run some benchmarks as well? (Feel free to choose, so long as it's something I can run on a modern system to make a direct comparison.)

These are systems that my family has had in the past... basically I'm wanting to figure out just how much of an uplift in performance we've had over the years of upgrades. (For example, how much faster is an R7 5950X, TR Pro 3995X or Epyc 7742 (or at least an i7-6700K / 4790K) vs a 286-10, with the same software / benchmark on both. )

Spoiler

2002 family desktop:
- 1.4 GHz Athlon CPU, Socket A, 266 FSB
- MSI K7T266 Pro2 Socket A motherboard
- 256MB DDR PC2100 RAM
- 40MB IBM IDE 7200rpm HDD
- Windows 2000
1999 family desktop:
- Pentium 166 MMX
- "VX up to 233 MHz" motherboard (idk what model exactly)
- 64MB RAM (one invoice hints that it may have involved "SIMM 4Mx32-60 72-pin")
- ATI All-In-Wonder (originally bought August 1997 - the original / 1st-generation AIW)
  - also Trident 9685 4MB
  - or Fahrenheit ProVideo 64
- 6.4GB Seagate IDE HDD
- Windows 98
  - On this one, I'd like the original Half-Life / Team Fortress Classic / Counter-Strike 1.6 benchmarked.
  - I think I remember getting < 9 fps at 320x240, software mode, in TFC on 2fort with my face buried in a wall or sniper scoped on the skybox, but I wonder if my memory is playing tricks on me and it wasn't quite that bad?
    - Side note: that's what I imagine Crysis ran like when it first came out, on the then-equivalent of 4-way SLI Titans / *TX **90s / Quadro *6000s / Tesla *100s / AMD counterparts, or maybe about 1/3-1/2 the fps. I never played it, though. My experience with TFC on this PC inspired my minimum standards for "But can it run Crysis / Game F?" - it doesn't have to get 60+fps at 1080p+ max settings, it just has to launch / run at all.
  - I also remember getting ~300-400+ ping on dial-up and playing pretty much exclusively as an engineer, relying mostly on my sentry gun, but I think it'd be a bit much to ask to benchmark gaming on dial-up.
1998/1999 brother's desktop (btw the "or" may have been an earlier PC to about 1997 or so):
- Pentium II 400 CPU
- QS440BX motherboard
  - or Tyan Tahoe Dual Pentium II
- 128MB SDRAM DIMM RAM
  - or 16MB EDO
- Gigabyte Riva TNT GPU
- 13.6GB Maxtor IDE HDD
- Windows 98
1995 family desktop:
- AMD (486) DX4-120 CPU
- 486 VLB w/128K Cache Motherboard
- 4MB or 8MB RAM (I think - "invoice says "1x36-70"
- "Super VGA 1024x768 w/1MB video card
  - or possibly the Trident or the Fahrenheit mentioned above, idk
- Sound Blaster AWE32 (or it may have been used with the 286 below in its later days, idk)
  - (also there was a Sound Blaster PCI64 that may have been in a later PC)
- 540MB IDE Connor HDD
- Windows 95
  - Is there any possibility Half-Life / TFC / CS might have run, albeit barely, on this? We got the above Pentium 166 system around the time they came out, but we may have still been using this when we got the game, idk.
1989 family desktop:
- 286-10 CPU
- 640K RAM
- EGA graphics
- 40MB Seagate MFM HDD
- MS-DOS 3.3

In my laptop (Clevo P750DM-G), I have an i7-6700K which, while it runs fine at 4 GHz, for some reason Cinebench and Intel XTU show it as a 3.5 GHz CPU. (See screenshot for an example.)

Spoiler

\

Interestingly, an older screenshot of Cinebench R15 shows it saying "@ 4.01 GHz"...

Spoiler

but a few more recent screenshots that includes R15 and a few others say "@ 3.51 GHz".

Spoiler

results on the multi-CB run after all have run simultaneously, albeit on a different day:

Spoiler

results after re-running each benchmark individually:

Spoiler

I had thought it could be because I had originally had an i3-6100 in the laptop's LGA 1151 socket when I got it in December 2015, then upgraded to the 6700K in November 2016, without reinstalling Windows or XTU.

But, there's a couple things possibly wrong with that evaluation:

The i3-6100 is a 3.7 GHz CPU, not 3.5 GHz
I found an older screenshot with the 6700K (see inside the spoiler) that showed Cinebench R15 properly reporting it as 4 GHz.

What could have caused Windows / the software to think my i7-6700K is a 3.5 GHz CPU (it still runs fine at 4 GHz, and I can OC up to about 4.5, or 4.6 with difficulty and just barely on occasion 4.7, limited by my laptop's cooling)?

And what can I do to correct it so it reports itself properly?

I have an transcoding test in progress on my i7-4790K using these settings in Handbrake.

Spoiler

So far, it seems to be running at approximately 1 hour to encode each minute of footage.

A previous time, I had done an encode with what I thought were max quality settings (at 4K), and I have a couple screenshots from that one showing it about 82.40% through a 2:36 video, with an elapsed time of 1:00:32:54, and 05:21:29 remaining.

Spoiler

Unfortunately the settings are obscured in that screenshot, and I don't have a record (either readily available, or at all) of the settings I used that time.

Another time I ran a 17-second video through the wringer, and it was showing these settings, with 18.30% progress, 48:07 elapsed and 3:28:39 remaining. (This one was more of a torture test, as I think I was transcoding to 8K.)

Spoiler

I also think I remember running a 4-minute video through Handbrake, and having it take 4 days to transcode it, with maximum settings (at 4K resolution) on H.265. (I don't have a screenshot of that one, though.)

I'm trying to figure out, though, what might have made the vast difference in encoding speeds?

Could there have been some subtle setting I'm forgetting about that made a huge difference? I think I remember using RF=0, placebo preset, keyint=1, 3840x2160 output resolution and H.265 on most of those. (The current one is using 10-bit H.265, I tried 12-bit with a short test and I couldn't open the resulting file in VLC, IIRC.)

Could a different version of Handbrake have made any difference?

I know I'm not getting any GPU acceleration with the current run, as 1 - I'm not using QuickSync, and 2 - I don't have my GTX 1060 3GB physically installed right now.

Could it be due to the difference in source and destination media?

This time, I'm reading from an external 500GB Samsung T5 SSD, and writing to a 1TB Samsung 970 Evo.

I don't have "reliable" / readily-accessible records / screenshots / photos, but I think with some of the previous tests (like 4 days to transcode 4 minutes, or 1 day 6 hours to transcode 2 minutes 36 seconds), I may have been reading from and writing to the same 7200rpm hard drive. But I thought the encode was so slow anyway that it wouldn't be fast enough to saturate the HDD, and the screenshots above barely show any HDD activity.

Along with the current Handbrake transcode test, I extracted the audio (as uncompressed WAVE files) from the same videos, totaling a little over an hour, then transcoded them to 320kbps mp3, q=0.

That transcode took about a minute.

Basically, when the time comes to upgrade from my 4790K (thinking November 25, 2022 or later, or when 2nd-gen DDR5 CPUs/mobos are out (so I can let the kinks get worked out with 1st-gen)),

I'd like my new CPU to be as fast an encoding video, as my old CPU is at encoding audio.

And, preferably the same or better price/performance uplift vs what my dad got when upgrading from the 286-10 to the 486 DX4-120.

Spoiler

The 286-10 invoice from January 4, 1989, doesn't mention the CPU by itself, but says the combo of CPU, board, 640k RAM, 1.2 floppy, 101 keyboard was $939. (doesn't specifically mention case/PSU but I imagine that price included it since it's not anywhere else.)
- Other things I've seen online hint that the 286 CPU by itself could have been anywhere from about $275 to $500 or so - could someone verify or correct that?
The AMD DX4-120 invoice from October 12, 1995, mentions $102 as the price for that CPU.
I bought my 4790K on January 7, 2015.
The same number of days later gets me to October 14, 2021, and I'm tentatively planning on waiting another year or more past that to upgrade.

When I do upgrade, though, I want a long period of forward upgrade compatibility - to the point that I anticipate replacing the 80+ Platinum/Titanium 10+-year-warranty PSU I'd get then at least once, maybe twice, before replacing the motherboard.
Basically I want to be able to upgrade one part at a time, basically replacing a part here and there once every 1/2 to 2 years, or the same part every 3 to 7 (10 or 12 or whatever the warranty is, for the PSU) years or so.
Having to remove a bunch of parts to replace the motherboard, because a new CPU would require it, would not count as "one part at a time" upgrades. (Removing & reinstalling a CPU cooler would be okay, if it's MUCH easier to do than my Hyper 212 Evo.)

Side note, related to video transcoding...

What would be a better codec & settings to use for transcoding video for editing / archiving?

Spoiler

I definitely want all I-frames for editing, and for archiving, I'd prefer essentially lossless (or at least something that no matter how much I edit and re-save, it won't degrade over multiple edits like most lossy formats I'm aware of would do).

(For unedited / original footage, I would likely just keep the original files as they came straight from the camera; although, if there's a way to "losslessly" transcode THOSE to all I-frames, but otherwise the same quality, I'd like to know, in case I ever decide to do that. My camera, a Panasonic FZ-1000, doesn't have an all-I mode, and I'm probably not upgrading to a new camera for at least several years or so - I may be eyeing something like the Sony A7s IV or V sometime....)

And for distribution, or for playback on other devices, is there a video codec that

is efficient (in terms of space used vs quality),

is relatively fast to encode (vs other codecs with similar quality, encoded on the same hardware)

has wide-ranging compatibility (from the newest software & devices all the way back to devices / software that had already been EOL'd before the ORIGINAL MPEG / AVI specs were invented, AND across all generations of products / software)

?

Also another side note related to encoding ...

Spoiler

a lot of my older mp3 (audio) files that I had encoded in VBR show the wrong playtime, often by a huge difference. For example, there might be a particular song that I know was, say, 4 minutes long, but the playback device / software says it's like 26 minutes or so. (I'm suspecting it has something to do with the first frame being silence or noise, therefore 8 or 32 kbps instead of the average ~110-140 kbps through the whole piece, and somehow that makes whatever determines what the playtime is, think it's much longer than it actually is.) Also sometimes the software / device will constantly "update" the total time while it's playing VBR files.

Is there anything I can do to fix that? (Including so that it displays properly on older devices / software that don't support some kind of VBR tag that was coming up during my Google searches a while ago) ... or would I need to transcode them to CBR (or another format entirely - would have to be compatible with pretty much everything though) to make sure the times are accurate?

BTW in both cases (audio or video, and I would also apply to photos as well) ... one of my requirements for choosing a codec / format is that it be compatible with pretty much everything. (At least, everything digital all the way back to, if not before, the 1980s or 1970s or so.... No, I don't expect to load something like an mp3 file on a reel-to-reel tape or a vinyl record, or an mp4 file on a reel-to-reel film or VHS tape. )

I wish the previews would show the spoilers collapsed, so I'd get a better idea of the length of the main parts of the post.

March 2, 2021

You did not buy a 4790K in 1995.... lol. You need a little proof reading

Any how, I have a 4000+ s939 processor 130nm clawhammer.

https://www.cpu-world.com/CPUs/K8/AMD-Athlon 64 4000%2B - ADA4000DEP5AS (ADA4000ASBOX).html

I do have a 4400+ Brisbane here. Would this suffice?

https://www.cpu-world.com/CPUs/K8/AMD-Athlon 64 X2 4400+ - ADO4400IAA5DU.html

I do not have a stock cooler or the board you are looking for. Memory, I gotta check the stash.

LMK

-rascal- · March 2, 2021

Hmm... I might have my old Athlon 64 X2 4400+ Brisbane (65W variant) with me in storage...or I sold it.

I actually never got rid my old Athlon 64 X2 system...yet...THAT I know has a Athlon 64 X2 6000+ Windsor (89W).

Last time I checked, it was running Windows Vista ... because the Acer OEM motherboard would not accept Windows 7

2GB X2 of Corsair XMS2 DDR2-800 memory.

Motherboard is some Acer / Foxconn 690G chipset based motherboard with the ATi Xpress X1250 iGPU.

March 2, 2021

Just now, -rascal- said:

Athlon 64 X2 6000+ Windsor (89W).

That's a nice chip. Can I have it?

Mister Woof · March 2, 2021

45 minutes ago, ShrimpBrime said:

You did not buy a 4790K in 1995.... lol. You need a little proof reading

Any how, I have a 4000+ s939 processor 130nm clawhammer.

https://www.cpu-world.com/CPUs/K8/AMD-Athlon 64 4000%2B - ADA4000DEP5AS (ADA4000ASBOX).html

I do have a 4400+ Brisbane here. Would this suffice?

https://www.cpu-world.com/CPUs/K8/AMD-Athlon 64 X2 4400+ - ADO4400IAA5DU.html

I do not have a stock cooler or the board you are looking for. Memory, I gotta check the stash.

LMK

for me 1995 was a 486DX/33

because we were poor

-rascal- · March 2, 2021

1 minute ago, ShrimpBrime said:

That's a nice chip. Can I have it?

Is it?

I didn't know better at the time, just needed a lower TDP replacement CPU, so the motherboard won't get angry.

I recall there is a 125W version of the Athlon 64 X2 6000+ ..... and another 89W based on the Brisbane..?

March 2, 2021

18 minutes ago, -rascal- said:

Is it?

I didn't know better at the time, just needed a lower TDP replacement CPU, so the motherboard won't get angry.

I recall there is a 125W version of the Athlon 64 X2 6000+ ..... and another 89W based on the Brisbane..?

The Windsor core chips clocked better than the 65nm Brisbanes. At least with any of my experiences.

I think the 125w version is a "Black Edition" But don't quote me on that. Most of mine where 95w and 65w chips (all different makes and models and platforms)

The last Windsor core I had, was a set of twins I got from E-Bay. Both sold cheap (25$) for the pair at the time, both missing pins and marked assumed non working.

Both worked. EDIT IN 6400+ here.

Both hit 4ghz and still holds 1st US to this day at 4063mhz. (Or somewhere close to that without looking)

Yeah, it's a good chip. If you get into tweaking it with a good board, probably hit 600mhz memory frequency on the right divider.

The Brisbane was a hit or miss OCer.

Some brick walled like the Agena Phenoms 3ghz, not a stitch more.

Suffered TLB errata which was later cured with the Phenom II.

_____TLDR - Skip the rest.______________

History (short)

The 5000+ Athlon ending in DGI (If I remember this all right tonight) was an unlocker.

When unlocked with LLC (or was it ACC?) the Cpu would post as an FX chip.

The FX name was dropped from the Phenom name and they used II in place of it.

Then we got the real FX chips and got raped by frequency marketing tactics lol.

Cpu-Z screen shot. Phenom FX

(I recently tossed this board too)

Spoiler

Edited March 2, 2021 by ShrimpBrime
Additional Comment

-rascal- · March 2, 2021

23 minutes ago, ShrimpBrime said:

The Windsor core chips clocked better than the 65nm Brisbanes. At least with any of my experiences.

I think the 125w version is a "Black Edition" But don't quote me on that. Most of mine where 95w and 65w chips (all different makes and models and platforms)

The last Windsor core I had, was a set of twins I got from E-Bay. Both sold cheap (25$) for the pair at the time, both missing pins and marked assumed non working.

Both worked. EDIT IN 6400+ here.

Both hit 4ghz and still holds 1st US to this day at 4063mhz. (Or somewhere close to that without looking)

Yeah, it's a good chip. If you get into tweaking it with a good board, probably hit 600mhz memory frequency on the right divider.

The Brisbane was a hit or miss OCer.

Some brick walled like the Agena Phenoms 3ghz, not a stitch more.

Suffered TLB errata which was later cured with the Phenom II.

_____TLDR - Skip the rest.______________

History (short)

The 5000+ Athlon ending in DGI (If I remember this all right tonight) was an unlocker.

When unlocked with LLC (or was it ACC?) the Cpu would post as an FX chip.

The FX name was dropped from the Phenom name and they used II in place of it.

Then we got the real FX chips and got raped by frequency marketing tactics lol.

Cpu-Z screen shot. Phenom FX

(I recently tossed this board too)

Reveal hidden contents

Shiiiieeet.

Blast from the past. The more you know hahah

Damn, didn't know they had the FX in the naming on Phenom CPUs.

I just remember the Athlon 64 FX chips from back then.

Now that I think of it, I guess that was another reason for AMD to bring back the FX naming back with 'Bulldozer' and 'Vishera' on socket AM3+. They used FX naming (briefly) with 1st Gen Phenon.

PianoPlayer88Key · March 3, 2021

1 hour ago, ShrimpBrime said:

You did not buy a 4790K in 1995.... lol. You need a little proof reading

Haha thanks yeah I missed that, had 1995 and the 486 on the mind Was actually 2015, I edited & corrected the post.

1 hour ago, ShrimpBrime said:

Any how, I have a 4000+ s939 processor 130nm clawhammer.

https://www.cpu-world.com/CPUs/K8/AMD-Athlon 64 4000%2B - ADA4000DEP5AS (ADA4000ASBOX).html

I do have a 4400+ Brisbane here. Would this suffice?

https://www.cpu-world.com/CPUs/K8/AMD-Athlon 64 X2 4400+ - ADO4400IAA5DU.html

I do not have a stock cooler or the board you are looking for. Memory, I gotta check the stash.

LMK

The 4400+ *might* work, if it could be downclocked to whatever the 4000+ ran at stock? (I think it was 2.1 GHz.)

29 minutes ago, -rascal- said:

Hmm... I might have my old Athlon 64 X2 4400+ Brisbane (65W variant) with me in storage...or I sold it.

I actually never got rid my old Athlon 64 X2 system...yet...THAT I know has a Athlon 64 X2 6000+ Windsor (89W).

Last time I checked, it was running Windows Vista ... because the Acer OEM motherboard would not accept Windows 7

2GB X2 of Corsair XMS2 DDR2-800 memory.

Motherboard is some Acer / Foxconn 690G chipset based motherboard with the ATi Xpress X1250 iGPU.

The 6000+ would be a bit more iffy though - if it could clock down to 2.1 GHz, maybe? (My 4000+ was the 65W / 65nm Brisbane variant.)

I'm basically trying to see if I can get a screenshot or few of a copy of my old CPU running benchmarks, at the same settings (stock) I would have used back in the day, had I run the benchmarks myself. (I couldn't run Cinebench R15, as I was on Windows XP Home 32-bit, never had a 64-bit OS on that PC.)

(It would help if it actually said the same parts, rather than 4400+ or 6000+, though.)

My dad has a Core 2 Duo laptop, that I've run a few benchmarks on (see spoiler), but I'd like to be able to compare my i7-4790K + HD 4600 / GTX 1060 3GB desktop and i7-6700K + 970M 6GB laptop to the system I used to have.

Spoiler

(the non-64-bit version scored 3746.)

Cloud Gate scored 0 overall with:

Graphics score = 0

G. Test 1 = 0.00 FPS, G. Test 2 = 0.43 FPS

Physics score = 781

P. Test = 2.48 FPS

Interesting, 3DMark 2001 got a much higher score under XP (the OS my dad had when he was using the laptop) than under 10 (I got a 240GB Crucial BX200 after he quit using the laptop, and stuck it in the laptop & put W10 on that).

Oh, while searching for a few screenshots, apparently I found this one from an Athlon 64 X2 4000+ Brisbane CPU .... but it only shows multi-core score (not single-core - I may have forgotten to mention in the OP that I also wanted that), and it's R15, not the newer versions (R20, R23) of Cinebench. (It shows R15 from my 4790K, as well as R20 from my dad's old C2D T7250 and my 6700K - I had put them side-by-side, had almost forgotten about this one.)

Spoiler

March 3, 2021

3 hours ago, PianoPlayer88Key said:

Haha thanks yeah I missed that, had 1995 and the 486 on the mind Was actually 2015, I edited & corrected the post.

The 4400+ *might* work, if it could be downclocked to whatever the 4000+ ran at stock? (I think it was 2.1 GHz.)

The 6000+ would be a bit more iffy though - if it could clock down to 2.1 GHz, maybe? (My 4000+ was the 65W / 65nm Brisbane variant.)

I'm basically trying to see if I can get a screenshot or few of a copy of my old CPU running benchmarks, at the same settings (stock) I would have used back in the day, had I run the benchmarks myself. (I couldn't run Cinebench R15, as I was on Windows XP Home 32-bit, never had a 64-bit OS on that PC.)

(It would help if it actually said the same parts, rather than 4400+ or 6000+, though.)

My dad has a Core 2 Duo laptop, that I've run a few benchmarks on (see spoiler), but I'd like to be able to compare my i7-4790K + HD 4600 / GTX 1060 3GB desktop and i7-6700K + 970M 6GB laptop to the system I used to have.

Reveal hidden contents

(the non-64-bit version scored 3746.)

Cloud Gate scored 0 overall with:

Graphics score = 0

G. Test 1 = 0.00 FPS, G. Test 2 = 0.43 FPS

Physics score = 781

P. Test = 2.48 FPS

Interesting, 3DMark 2001 got a much higher score under XP (the OS my dad had when he was using the laptop) than under 10 (I got a 240GB Crucial BX200 after he quit using the laptop, and stuck it in the laptop & put W10 on that).

Oh, while searching for a few screenshots, apparently I found this one from an Athlon 64 X2 4000+ Brisbane CPU .... but it only shows multi-core score (not single-core - I may have forgotten to mention in the OP that I also wanted that), and it's R15, not the newer versions (R20, R23) of Cinebench. (It shows R15 from my 4790K, as well as R20 from my dad's old C2D T7250 and my 6700K - I had put them side-by-side, had almost forgotten about this one.)

Reveal hidden contents