VOGONS


Christmas GPU Upgrades, benchmark suggestions?

Topic actions

First post, by clueless1

User metadata
Rank l33t
Rank
l33t

My son's i5-4570S is going from a GTX 750Ti to a GTX 1060. My XP system with a C2D 2.67Ghz is going from 8800GTX to GTX 750Ti. I'd like to start gathering some benchmark numbers with the current cards, then run same benchmarks on the new cards when they are put in on Christmas. 😀

Suggestions?

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 1 of 20, by ODwilly

User metadata
Rank l33t
Rank
l33t

Crysis demo!

Main pc: Asus ROG 17. R9 5900HX, RTX 3070m, 16gb ddr4 3200, 1tb NVME.
Retro PC: Soyo P4S Dragon, 3gb ddr 266, 120gb Maxtor, Geforce Fx 5950 Ultra, SB Live! 5.1

Reply 2 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t
ODwilly wrote:

Crysis demo!

I assume for the XP rig. Thanks, I'll look for it. I assume it has a built-in benchmark? The most demanding game I think on my son's PC is COD: Black Ops III. Looking now for a way to benchmark it. I've got the new 3dmark demo off of Steam, so I'll do that too.

edit: looks like Just Cause 2 has a built-in benchmark, and he's got that game. Cool 😀

edit 2: downloading unigine valley (runs on XP!) and crysis demo now.

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 3 of 20, by Jade Falcon

User metadata
Rank BANNED
Rank
BANNED

3dmark 2003, 2005 and 2006.

If you follow hwbot rules I'll post the scores on the vogons team account if you'd like.

Reply 4 of 20, by Standard Def Steve

User metadata
Rank Oldbie
Rank
Oldbie
Jade Falcon wrote:

3dmark 2003, 2005 and 2006.

If you follow hwbot rules I'll post the scores on the vogons team account if you'd like.

Umm, those are going to be completely CPU limited with the 1060.

If you gotta run 3DMark, use the latest version. That'll let the new GPU stretch its legs.

"A little sign-in here, a touch of WiFi there..."

Reply 5 of 20, by agent_x007

User metadata
Rank Oldbie
Rank
Oldbie

I guess build-in free games only... can you post a list of games available to you ?

Crysis Benchmark Tool (should work on demo version, make sure you update it to 1.2 version if possible).
Other games : Half Life 2 : Lost Coast ?

Unigine Valley is great GPU stress test.
3DMark 11 ("P" test) + Fire Strike and/or Sky Diver, should do the trick as well (for GTX 1060 at least).

You got Doom (2016) ?
There is a stats option you can enable to see FPS and CPU/GPU load real-time.

Also, remeber that you can switch Steam accounts on one PC (to unlock new one however, you will need a passcode which will be send on email account you connected to new steam account).

157143230295.png

Reply 6 of 20, by kanecvr

User metadata
Rank Oldbie
Rank
Oldbie
agent_x007 wrote:
I guess build-in free games only... can you post a list of games available to you ? […]
Show full quote

I guess build-in free games only... can you post a list of games available to you ?

Crysis Benchmark Tool (should work on demo version, make sure you update it to 1.2 version if possible).
Other games : Half Life 2 : Lost Coast ?

Unigine Valley is great GPU stress test.
3DMark 11 ("P" test) + Fire Strike and/or Sky Diver, should do the trick as well (for GTX 1060 at least).

You got Doom (2016) ?
There is a stats option you can enable to see FPS and CPU/GPU load real-time.

Also, remeber that you can switch Steam accounts on one PC (to unlock new one however, you will need a passcode which will be send on email account you connected to new steam account).

^this.

I generally use 3dmark fire strike, Unigine Valley, Unigine Heaven and more recently I've been using Doom 2016 - these cover direct 3d, openGL and Vulkan performance.

The rest of the 3dmark suite (sans 2011) is too old to measure performance adequately.

A little off-topic, but I'd have gone for the RX 480 over the GTX 1060. While the 1060 is 2-10% faster in DX11, the RX 480 is a lot faster in DX12 and Vulkan, making it better for future games. I've also been dissapointed with nvidia drivers recently - there was one driver that messed up video card fan controls, and another more recent one that messed up video memory clocks (you can google it). Also considering the gameworks bullshit and nerfing of previous generation cards, I'd rather give my money to AMD.

Reply 7 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t

@kanecvr, yeah, it was a tough call. I kept going back and forth between the 1060 and RX 480. I'm not sure AMD's lead in DX12 and Vulcan is big enough to practically extend the life of the video card in that system. You know? By the time there are games out that might showcase enough of a performance gap between the two, they'll both be too slow to be playable and I'd be better off upgrading anyway. So it came down to the 1060's power/heat efficiency advantage, which will be there for the life of the card. But yeah, I went back and forth A LOT trying to decide. It didn't hurt that I got the GTX 1060 for $225 shipped on Black Friday. 😀

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 8 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t

Got some results for the C2D/8800GTX on WinXP. A good mix of old and newish. Now to work on some i5/750Ti results. Will probably use Unigine Valley, new 3DMark, not sure what else. Maybe Crysis.

The attachment 8800vs750.jpg is no longer available

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 9 of 20, by awgamer

User metadata
Rank Oldbie
Rank
Oldbie
clueless1 wrote:

My son's i5-4570S is going from a GTX 750Ti to a GTX 1060. My XP system with a C2D 2.67Ghz is going from 8800GTX to GTX 750Ti. I'd like to start gathering some benchmark numbers with the current cards, then run same benchmarks on the new cards when they are put in on Christmas. 😀

Suggestions?

Games you and your son play.

Reply 10 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t
awgamer wrote:
clueless1 wrote:

My son's i5-4570S is going from a GTX 750Ti to a GTX 1060. My XP system with a C2D 2.67Ghz is going from 8800GTX to GTX 750Ti. I'd like to start gathering some benchmark numbers with the current cards, then run same benchmarks on the new cards when they are put in on Christmas. 😀

Suggestions?

Games you and your son play.

I play older FPS games on mine (Doom3, Far Cry, Crysis, etc) and RPGs (like Legend of Grimrock). I won't benefit much from the GTX 750Ti, other than higher details in something like Crysis. But I will use less power and generate less heat. 😀 Son's big game now is COD: BO3, but he plays some Minecraft, Rising Storm, and some other modern FPS games.

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 11 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t

Thanks for the suggestions. This is what I settled on.
On my XP machine that is going from 8800GTX to 750Ti:
Far Cry
Doom3
F.E.A.R.
3DMark06
Unigine Valley
Crysis Demo
This system runs at 1280x1024, and I'll be maxing everything out on both cards. You can already see the 8800GTX results in my last post above.

On the Win10 PC that's going from the 750Ti to the GTX 1060:
Far Cry
Doom3 (if it runs, last time I tried on this system it would not launch)
F.E.A.R.
Unigine Valley
Crysis Demo
Just Cause 2
new 3DMark
This system runs at 1680x1050. True 1080p monitor will have to wait for birthday or next Christmas.

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 12 of 20, by awgamer

User metadata
Rank Oldbie
Rank
Oldbie

>This system runs at 1680x1050. True 1080p monitor will have to wait for birthday or next Christmas.

Check craigslist, where I'm at 23-24" 1080p lcds go for $20-40. It's how I got my current monitor, 22" 1680x1050 for $30(going rate for used lcds have dropped since when I bought mine,) it's a TN panel but it doesn't have any input lag, which was the one feature I would check so you do don't end up with a too old lcd with lag that's bad for gaming. If TN is below you:), search for ips, much rarer but they pop up for sale.

Reply 13 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t

Got the GTX 1060 in my son's PC. He's busy playing COD: BO3, so I put his old GTX 750Ti into my XP rig. Will take benches from his system when he takes a break. Meanwhile, here's how the 750Ti compares to the 8800GTX it replaced. The CPU is a 2.67Ghz C2D, so there's plenty of CPU limiting going on, but overall, a great upgrade.

The attachment 8800vs750-final.jpg is no longer available

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 14 of 20, by Malik

User metadata
Rank l33t
Rank
l33t

It'll be more interesting to see what happens when you put the 1060 in YOUR system and bench the same set again to see how far you can go, despite the C2D limitations.

It's good to see the 8800GTX is no slouch and is not too far behind the 750Ti, a respectable performer for an older generation card. But I guess the power consumption / efficiency alone is worth the upgrade to a newer generation Nvidia card.

5476332566_7480a12517_t.jpgSB Dos Drivers

Reply 15 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t
Malik wrote:

It'll be more interesting to see what happens when you put the 1060 in YOUR system and bench the same set again to see how far you can go, despite the C2D limitations.

It's good to see the 8800GTX is no slouch and is not too far behind the 750Ti, a respectable performer for an older generation card. But I guess the power consumption / efficiency alone is worth the upgrade to a newer generation Nvidia card.

As far as I know the 1060 does not have drivers for XP, so I don't think that's going to happen.

If you look at the right numbers, you can see that without CPU bottlenecks, the 750Ti is about 2.5x faster than the 8800GTX. Don't get me wrong, the 8800GTX is a great performer (and better balanced with the rest of the system), but...130W just sitting at the desktop? That's crazy talk.

BTW, I got one benchmark result from the 1060: 3DMark Firestrike 1.1. The system is an i5-4570S, 2.9Ghz base frequency, 3.6Ghz turbo frequency. The only overclocking I've done is not really an overclock. There's a setting in the BIOS to force all 4 cores to run at 3.6Ghz during turbo mode that I've turned on. The GTX 750Ti got 4083 in the benchmark. The GTX 1060 gets 9927, roughly 2.4x faster. Interesting that the speed-up from 750->1060 is roughly the same as 8800->750.

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 16 of 20, by agent_x007

User metadata
Rank Oldbie
Rank
Oldbie

8800 GTX doesn't have a 2D/Low 3D clock option.
Basicly : It's always running at max. clocks (that's why Idle power draw and temps are so high).

As for results :
I think Crysis demo with AA score is wrong for GTX 750 Ti... ("free" 8x AA on GTX 750 Ti - that's simply not possible).
Could you check how 8xQ (or 4x) options looks like FPS wise on GTX 750 Ti (no AA vs. w/AA)?

FYI : Crysis AA "8x" option means MSAA 4x with additional cover samples (newer cards doesn't support this).
8xQ is true MSAA 8x.

157143230295.png

Reply 17 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t
agent_x007 wrote:
I think Crysis demo with AA score is wrong for GTX 750 Ti... ("free" 8x AA on GTX 750 Ti - that's simply not possible). Could yo […]
Show full quote

I think Crysis demo with AA score is wrong for GTX 750 Ti... ("free" 8x AA on GTX 750 Ti - that's simply not possible).
Could you check how 8xQ (or 4x) options looks like FPS wise on GTX 750 Ti (no AA vs. w/AA)?

FYI : Crysis AA "8x" option means MSAA 4x with additional cover samples (newer cards doesn't support this).
8xQ is true MSAA 8x.

You're right, thanks for catching that. I did some testing with the 750Ti (I'm not going to put the 8800GTX back in). It seems that if I run 8x first, it gives the "too fast" results I posted. If I run 8xQ, I get extremely slow results:

Play Time: 176.26s, Average FPS: 11.35
Min FPS: 8.65 at frame 460, Max FPS: 13.34 at frame 1760

Then if I go back and run 8x after running 8xQ, 8x results look right (slightly faster than 8xQ). Here's the GPU demo at 4x:

!TimeDemo Run 2 Finished.
Play Time: 43.67s, Average FPS: 45.80
Min FPS: 25.78 at frame 139, Max FPS: 67.96 at frame 88

So it seems the 8x results for both cards should be thrown out from my chart above.

EDIT: What about the 8x AA results for Unigine Valley on the 750Ti? Those are nearly "free" due to CPU bottlenecking.

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 18 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t

And here are results of 750Ti vs GTX 1060 on the Win10 i5 machine. I threw out the Crysis numbers. I had similar issues that I did above, plus I noticed that the results would fluctuate quite a bit between runs. The Far Cry benchmark app would crash every time I tried to run the benchmark with the GTX 1060.

The attachment 750vs1060.jpg is no longer available

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 19 of 20, by clueless1

User metadata
Rank l33t
Rank
l33t

BTW, one of the issues I had with considering the 1060 is I could not find any with two DVI connectors. We have two old monitors, with only VGA and DVI outputs. My solution, since the 1060 we got has only one DVI, is to use the bigger monitor with the 1060, and the smaller monitor (1440x900) with the integrated HD4600. I had to change a BIOS setting to make the integrated graphics "primary" in order for this to work, but it's a pretty seamless solution.

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks