• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

~$200 GPU

B

Bland_JamesBland

Weaksauce
Joined
Aug 29, 2013
Messages
86
My girlfriend wants to get me a birthday gift and she can't come up with anything herself claiming I'm to difficult to buy for. lol So I said I'd like a GPU for my pc. She agreed and said tell me which one...

I've got an A10-7850k, 8gb 2133Mghz ram, 1 ssd, 1 hdd, seidon 120v with 2 120mm fans. and a 500w psu...
GTX960 or R9 380?

Thanks!
 
380. Based on Tonga. Look at the bundles too (games) that they come with and if either strike your fancy. 380 comes with Dirt Rally where the 960 has something with Heros of the Storm.
 
Last edited:
I think the 380 might be better. But my fiancee loves her 960 for all her games. She's only running an i5-4490 in it but all the games she throws at it run great and she's happy. Plus it sips power.
 
The R9 380 is a poor card.

Better to get the GTX 960. Nvidia drivers have better scaling to slower CPUs like yours (they can use the extra cores).

http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/21

Notice how, aside from corner cases like Shadow of Mordor, the Nvidia GPU gives much higher average frame rates on the AMD A10-7850k?

And while the GTX 770 is FASTER than the R9 285, you'll notice that at no time is the R9 285 GPU-limited on your A10-7850k (it gets faster when running on Intel). Thus, the difference in CPU efficieicny between Nvidia and AMD is real.

This should go away with DirectX 12, but those titles are a long way off. Buy for the games you play today :D

Also, the R9 380 has exactly the same performance as the GTX 960 when BOTH cards are fully overclocked:

http://www.hardocp.com/article/2015/08/04/msi_geforce_gtx_960_gaming_4g_video_card_review/9

I'm amazed people here continue to recommend the R9 380, despite the poor drivers, and high power consumption!
 
Last edited:
defaultluser said:
The R9 380 is a poor card.

Better to get the GTX 960. Nvidia drivers have better scaling to slower CPUs like yours (they can use the extra cores).

http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/21


I'm amazed people here continue to recommend the R9 380, despite the poor drivers, and high power consumption!
Click to expand...

The 770 actually had a larger % (2.97% for the 285 and 4.15% for the 770) decrease than the 285 in your link, comparing a 4790k vs. 7850 (OPs CPU). Suggesting your statement that nVidia drivers are better isn't true.

AMD has fine single card drivers. Multicard sucks for both vendors.

If 60 watts is a factor for the 380 then that is what it is. If you game 4 hrs/day every day it's about $9 a year at $0.10/kWh more cost against the 380.
 
Last edited:
Dayaks said:
The 770 actually had a larger % (2.97% for the 285 and 4.15% for the 770) decrease than the 285 in your link, comparing a 4790k vs. 7850 (OPs CPU). Suggesting your statement that nVidia drivers are better isn't true.

AMD has fine single card drivers. Multicard sucks for both vendors.

If 60 watts is a factor for the 380 then that is what it is. If you game 4 hrs/day every day it's about $9 a year at $0.10/kWh more cost against the 380.
Click to expand...

Project Cars says no to your thesis. AMD drivers have huge (compared to NVIDIA) CPU overhead.
 
Dayaks said:
The 770 actually had a larger % (2.97% for the 285 and 4.15% for the 770) decrease than the 285 in your link, comparing a 4790k vs. 7850 (OPs CPU).
Click to expand...

Why are you comparing the 4790k versus the 7850?

My intent was for you to compare the 7850 scores between the two cards, in the cases where we were not GPU-limited.

The 4790 scores are only useful to let you know when a test is GPU-limited or not. If the framerates are significantly higher on the 4790 , then we are CPU-limited on the 7850. Does that not malke sense to you?

trandoanhung1991 said:
Project Cars says no to your thesis. AMD drivers have huge (compared to NVIDIA) CPU overhead.
Click to expand...

Right

Two of the games in my link are GPU-limited (alien Isolation, Shadow of Mordor). You can't use those benchmarks to determine anything.

HOWEVER, ,in the three games where we are NOT GPU-limited, Nvidia shows a marked improvement:

The Total War framerates are 35% higher 19.1 versus 14.4.

The GTA V framerates are 25% faster: 42.5 versus 33.8, used 7870k.

GRID is 227% faster: 102 versus 45!

And in the high-end test of a 290x versus a 980, the Alien Isolation benchmark is no-longer GPU-limited, and it posts a 10% faster framerate.

http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/22

Shadow of Mordor is still the same on Nvidia as it is on AMD in these non-GPU limited tests, but the numbers (4 games to 1) are not in AMD's favor. When you have a weaker CPU, how badly do you want to risk that you're playing one of AMD's rare optimized games? And BEST CASE, AMD scales the same as Nvidia.

And I noticed NOBODY CONTRADICTED my claim about the GTX 960 and the 380 performing the same when both are overclocked! So now you have absolutely no ground on which to recommend new AMD at $200!
 
Last edited:
For 200 your best bet is to score a used 290X graphics card. Also if she truly loves you, ask her to buy you a new 290X which usually go for 250/260 bucks on newegg or Amazon depending on what time of day you are looking at.
 
Ha my phone only loaded a quarter of the page for whatever reason. I didn't get to the good results.

I am not sure if you understand what you are saying. You ignore the fact the 770 is a much faster card and quote % differences. You mention GTA V but the % decrease between the faster and slower CPUs are similar between the two cards suggesting no additional overhead.

Let's look at your last example: Alien Isolation 290x vs 980... You are using a game that is getting 182 vs 137 with a good CPU then comparing GPU to GPU on a 7850 (122 vs 133) and saying ha! nVidia does better CPU constrained! It doesn't work that way. The GPU still matters. If they started out both at 182 and then AMD fell further I'd say sure. Even if the CPU is the bottle neck it doesn't completely negate the video card. A Titan X (and perhaps Fury X) would get slightly above 133. A worse card would score lower. To illustrate my point using the anandtech reviews you linked, compare the 770 and 980 on GTA V. Really CPU limited with a shitty CPU like the 7800. The 770 39.99 FPS and the 980 48.43 FPS. CPU bottle neck, GPU still matters.

No one contradicted that the 960 and 380 perform similarly because they do. They are also have similar pricing, both have bundles, one bundle might be more preferable.

If I was you I would have just pointed to the Grid results with the 7850 and left it at that. Legimate driver issues there. :D

Honestly the best advice is getting a used 290/290x if you are into that. Have to pay more attention and get a good model though.
 
Last edited:
GTX 960 so your computer doesn't become a furnace unless you want one. For example I went from a 750 ti to a HD 7950; my room became noticeablely hotter and it became so uncomfortable that I had to keep my door and my windows open to stay cool (rarely use AC). If you can take the heat then r9 380.
 
Last edited:
vividshock said:
GTX 960 so your computer doesn't become a furnace unless you want one. For example I went from a 750 ti to a HD 7950; my room became noticeablely hotter and it became so uncomfortable that I had to keep my door and my windows open to stay cool (rarely use AC). If you can take the heat then r9 380.
Click to expand...

The 7950 is also 50% faster than a 750ti. The extra heat we're talking about is like a 60W light bulb... or 4% of a space heater with regards to the 380 vs 960.
 
Okay, looking closer at the numbers, they're not consistent enough between the two tests.

I wish they had not mixed older Kepler with Maxwell, and Hawaii with Tonga. More consistent architectures, and more than five games would have actually been conclusive.

It used to be before DirectX 10 that if you were CPU-limited in a game, the faster GPU made no difference. But the new fully-programmable system means you can have vastly different CPU scaling from one architecture to another, varying from one game to another! It makes it much harder to get a feel for CPU efficiency.
 
Last edited:
Between the 2 I'd go for the R9 380
Click to expand...

+1. I've used both, 960 and 380 (actually, an R9 285, essentially same thing as an R9 380).

In my experience, the 285 was definitely faster on average, and power and heat really weren't a concern. Not a noticeable difference for me, and may be worth considering only if you're using an ITX case, or are PSU-limited. The most recent AMD drivers worked great.
 
Depends on what games the OP intends to play, most likely the Kaveri chip will hold back either card. I have a 7700K on an Asus A88X-Pro and an Athlon 860K on a Gigabyte UP4, neither of them shows very good performance with my 380's even with a heavy overclock. My current game, Dragon Age Inquisition, is nearly unplayable with either chip with frames in the 20's regardless of settings and 95%+ usage on all four cores. Mantle isn't an option because it increases load times from a few seconds to something you could measure tectonic plate shift with.
 
defaultluser said:
Okay, looking closer at the numbers, they're not consistent enough between the two tests.

I wish they had not mixed older Kepler with Maxwell, and Hawaii with Tonga. More consistent architectures, and more than five games would have actually been conclusive.

It used to be before DirectX 10 that if you were CPU-limited in a game, the faster GPU made no difference. But the new fully-programmable system means you can have vastly different CPU scaling from one architecture to another, varying from one game to another! It makes it much harder to get a feel for CPU efficiency.
Click to expand...

I second that the GPU choices were strange to me.

nVidia is more consistent overall. Single GPU AMD is generally fine too, one odd game here and there causes issues and performance tanks (Project Cars). One problem with some sites like anandtech is I am pretty sure they don't retest like [H] does. A lot of games are vastly improved a day or week after launch (sadly). Brent usually revisits them like Farcry 4. Ensures same, up to date drivers ect. I like anandtech's data because it's so vast but the 7850 was probably tested on much older drivers and such.

Usually games are fixed quite quickly after launch for either vendor. Maybe I am getting older but I don't mind waiting - I took a few month break from the Witcher 3 for the all bugs to be fixed.
 
Last edited:
I'm somewhat confused. How is the GTX 770 and 960 related in performance? I thought the 770 was a tier above it at least. I could be wrong though. ;) And if he's spending $240 then he should get a R9 290 (new) or GTX 970 (new) or a used r9 290x. I thought his budget was $200 though.
 
cageymaru said:
I'm somewhat confused. How is the GTX 770 and 960 related in performance? I thought the 770 was a tier above it at least. I could be wrong though. ;) And if he's spending $240 then he should get a R9 290 (new) or GTX 970 (new) or a used r9 290x. I thought his budget was $200 though.
Click to expand...

It's the only review site I've ever found that actually tests multiple games with different GPUs on different CPUs. Most of the rest of the internet seems to be incapable of such a complex through process, so I had to take what I could get :D

Or did I miss a website you know about?
 
New: Go with the 960. It's newer tech than the 770 / 280X / 380 rebadges, and nearly as fast with far better thermals than any of the older options. You basically need a 970 or 290/390 to be in the next grade of performance. At that point, you're waaay over-budget.

Used: R9 290 (R9 390). Blows any of your "new" options out of the water (at the $200 range) and puts you up in the $300+ performance range for a cool hundred+ less.
 
As an eBay Associate, HardForum may earn from qualifying purchases.
It looks like the 770 is $300+, 960 can be as low as $190, 380 as low as $170.

What this guy said:
atp1916 said:
New: Go with the 960. It's newer tech than the 770 / 280X / 380 rebadges, and nearly as fast with far better thermals than any of the older options. You basically need a 970 or 390 to be in the next grade of performance.

Used: R9 290 (effectively a R9 390). Blows any of your "new" options out of the water (at the $200 range) and puts you up in the $300+ performance range for a cool hundred+ less.
Click to expand...
 
atp1916 said:
New: Go with the 960. It's newer tech than the 770 / 280X / 380 rebadges, and nearly as fast with far better thermals than any of the older options. You basically need a 970 or 290/390 to be in the next grade of performance. At that point, you're waaay over-budget.

Used: R9 290 (R9 390). Blows any of your "new" options out of the water (at the $200 range) and puts you up in the $300+ performance range for a cool hundred+ less.
Click to expand...

Maxwell and Tonga are about the same age and the 960 is not any faster than a 380 unless clocked to the moon. A used 290 is a way batter card obviously, but he isn't going to get much out of it with a Kaveri APU running the show.
 
As an eBay Associate, HardForum may earn from qualifying purchases.
defaultluser said:
It's the only review site I've ever found that actually tests multiple games with different GPUs on different CPUs. Most of the rest of the internet seems to be incapable of such a complex through process, so I had to take what I could get :D

Or did I miss a website you know about?
Click to expand...

Here ya go. :) This website will blow your mind with the amount of information on it. Use Google Chrome and allow it to translate for you. Bing actually has a better translator, (scary right?) but Microsoft didn't integrate it into MS Edge for some reason unknown to man.
http://udteam.tistory.com/793
 
cageymaru said:
Here ya go. :) This website will blow your mind with the amount of information on it. Use Google Chrome and allow it to translate for you. Bing actually has a better translator, (scary right?) but Microsoft didn't integrate it into MS Edge for some reason unknown to man.
http://udteam.tistory.com/793
Click to expand...

Holy shit, why have you been hiding this?
 
Here's is what I've read/seen in regards to the 960 vs 380.

960 has less CPU overhead (which is nice for an older i5 like mine)

380 is 5-ish percent faster on paper. But the min. FPS is far worse.

Then there's the driver issue. A new driver every six-ish months vs. monthly updates.

Unless someone can really give me a real, legitimate reason not to, I intend to purchase a heavily factory oc'ed 960 4gb in the next two weeks.
 
horrorshow said:
Here's is what I've read/seen in regards to the 960 vs 380.

960 has less CPU overhead (which is nice for an older i5 like mine)

380 is 5-ish percent faster on paper. But the min. FPS is far worse.

Then there's the driver issue. A new driver every six-ish months vs. monthly updates.

Unless someone can really give me a real, legitimate reason not to, I intend to purchase a heavily factory oc'ed 960 4gb in the next two weeks.
Click to expand...

Buying factory OC'd cards is a waste of money. Just buy a card with a nice cooler and overclock it. Nvidia cards do not run at a set speed. They power up and down according to the scene. Buying an overclocked card is pointless as it will vary in speed according to the demands of the scene.

AMD cards also vary in speed according to temperature. I set my speed to a static setting in MSI Afterburner on day one so no idea how it ramps up and down normally. It will ramp down if the temps get too high still.

So in short buy a stock card with a nice cooler.
 
I would agree with some posters that if you are okay with buying a used 290X etc, that route will get you the most performance for $200.
 
You must log in or register to reply here.
Back
Top