Yay for power efficiency

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Nhirlathothep

Senior member
Aug 23, 2014
478
2
46
www.youtube.com
I disagree with this statement completely. I've been following the GPU landscape for close to 15 years and I've noticed that the low powered videocards with good perf/watt tend to have large premiums and generally poor/awful performance/$. That means saving $30-50 on a high quality PSU actually ends up costing MORE long-term in terms of GPU upgrades. Cards such as GTX750/750Ti/950/960 all come to mind that had excellent performance/watt but horrible price/performance compared to much faster card that required a solid 500-550W PSU unit.

There have been so many deals on good PSUs over the years in the US, that there is 0, and I repeat 0 excuses why someone should sacrifice the choice of what CPU/GPU they can buy over the next 5-10 years to save what $20-50? Saving on the PSU is literally THE worst thing any new PC builder can do long-term. The PSU is the most important component in a system and money should first be saved on the motherboard, case, memory or just about anywhere else, not the PSU.

There is no reason a budget build cannot include a good PSU that can power any single GPU on the market. Some deals in the past:

XFX TS Series 750W Quiet ATX 80 Plus Bronze Certified Power Supply (P1-750S-NLB9) 5 Years Warranty $35 ( MIR $35)

750W Thermaltake Toughpower 80+ Gold Modular Power Supply (PS-TPD-0750MPCGUS-1) for $79.99 - $25 Rebate = $54.99.

EVGA 600W 80+ Certified ATX Power Supply (100-W1-0600-K1) on sale for $29.99.

700 Watt EVGA 700B 80 Plus Bronze Certified Power Supply (100-B1-0700-K1) [newegg.com]
The PSU also allows overclocking headroom and stability for both the CPU/GPU and upgrade path to more powerful 6-10 core CPUs. Furthermore, gamers who are just entering the PC gaming scene could eventually choose to move away from 6/8-pin videocards towards 200-250W GPU class of flagship cards. This is especially true as 3440x1440, 4K, 1440p 144-200Hz, etc. start to gain traction over the next 5 years.

I've been on the forums for so long that I can say without a shadow of a doubt any PC build with a budget/lower end PSU is an incorrectly build PC system. Any small amount of $$$ saved to meet some arbitrary budget will end up costing a lot more in the future.
I disagree
 

[DHT]Osiris

Lifer
Dec 15, 2015
15,289
13,587
146
Agree 110% with RussianSensation. I've had exactly two power supplies over the last decade and some change, each cost me about $100, were high quality (reviewed) when I bought them, and each is still in commission. One went from a Gen1 Core2Quad and an 8800GTX to an i7-950 with a 960gtx, other went from aforementioned i7 and a ... 475gtx? to currently a 6700k with a 970gtx. I've shouted from the rooftops to everyone that will listen, do not skimp on your base components. If you spend on the PSU and motherboard (quality! make sure you aren't spending on crap!), you will solve yourself hours and days of headaches and troubleshooting in the long run.

Seriously, I'd skimp on the video card before skimping on the PSU/mobo.
 

sm625

Diamond Member
May 6, 2011
8,172
137
106
I think the latest rumours had the GP107 featuring a single GPC or the equivalent of 640 CUDA cores. This would then be exactly half of a 1060 (GP106) and if the performance is also half of a 1060 , then it would be somewhere between a 950 and 960, not between a 960 and a 970.

I am thinking 896 cores, 120 sq mm, 128 bit. $150. $179 FE.
 

Nhirlathothep

Senior member
Aug 23, 2014
478
2
46
www.youtube.com
Agree 110% with RussianSensation. I've had exactly two power supplies over the last decade and some change, each cost me about $100, were high quality (reviewed) when I bought them, and each is still in commission. One went from a Gen1 Core2Quad and an 8800GTX to an i7-950 with a 960gtx, other went from aforementioned i7 and a ... 475gtx? to currently a 6700k with a 970gtx. I've shouted from the rooftops to everyone that will listen, do not skimp on your base components. If you spend on the PSU and motherboard (quality! make sure you aren't spending on crap!), you will solve yourself hours and days of headaches and troubleshooting in the long run.

Seriously, I'd skimp on the video card before skimping on the PSU/mobo.
This only means it s good for you.
When you assemble pcs to sell or for other pple, it s different
 

[DHT]Osiris

Lifer
Dec 15, 2015
15,289
13,587
146
It's even more important when you assemble for others. Tell them to put up or shut up, or you're gonna get a lot of phone calls for intermittent crap, because you built them a shoddy product. Don't just build a piece of garbage because it's all they can afford.
 
Reactions: DisarmedDespot

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com
How do you know the GP107 will have 33% less ROP than a GP106? If anything it might actually have less than half the ROPs of GP106, given that GP106 has an unusually high ROP count compared to GP104 and GP102 (50% higher relative to shader count). If GP107 only has a 33% lower ROP count than GP106 (and thus has 32 ROPs), then it would have double the relative ROP to shader amount of GP104 and GP102 (1 ROP per 20 CUDA cores, compared to 1 ROP per 40 CUDA cores for the GP104 and GP102).

It is true that if the GP106 is primarily bandwidth limited then the GP107 might trend closer to the 33% gap in bandwidth rather than the 50% gap in shader performance, but I'm not sure if there's any indications that the GP106 is more bandwidth limited than shader limited.

Every X07 card from Nvidia has had a 128-bit bus. Since after GM107, every card Nvidia has made has bundled 8 ROPs per 32-bit memory bus (128/32 x 8 = 32 ROPS). Further, GP106 is not that unusual in it's ROP count relative to the larger Pascal dies;it's only unusual compared to Maxwell. GP106 has exactly the same ratio of bus and ROPs as Kepler's GK106 had compared to GK104 and GK110. So I think I'm pretty spot on with my prediction of GP107 having 1 GPC, 128-bit memory bus, and 32 ROPs. The vram speed would only need to be 6ghz (vs. GP106's 8ghz) to be at half the bandwidth, saving some power in that regard and allowing Nvidia to slightly up-clock the card vs. GP106's core speeds to bring it in at ~55-58% GTX1060 performance at 65 watts power draw.

Assuming I'm right, vs. AMD's RX 460, GP107 would have both a larger deficit in DX12 and larger advantage in DX11 games (when compared to RX 480 vs. GTX 1060) based on the few reviews I've read, making the choice between the two cards all the more difficult.
 

antihelten

Golden Member
Feb 2, 2012
1,764
274
126
I am thinking 896 cores, 120 sq mm, 128 bit. $150. $179 FE.

896 cores would entail 7 SMs (128 cores per SM) which seems rather unlikely when Nvidia arranges their GPCs with 5 SMs each.

Most likely it will be 640 cores (5 SMs, 1 GPC).

Every X07 card from Nvidia has had a 128-bit bus. Since after GM107, every card Nvidia has made has bundled 8 ROPs per 32-bit memory bus (128/32 x 8 = 32 ROPS). Further, GP106 is not that unusual in it's ROP count relative to the larger Pascal dies;it's only unusual compared to Maxwell. GP106 has exactly the same ratio of bus and ROPs as Kepler's GK106 had compared to GK104 and GK110. So I think I'm pretty spot on with my prediction of GP107 having 1 GPC, 128-bit memory bus, and 32 ROPs. The vram speed would only need to be 6ghz (vs. GP106's 8ghz) to be at half the bandwidth, saving some power in that regard and allowing Nvidia to slightly up-clock the card vs. GP106's core speeds to bring it in at ~55-58% GTX1060 performance at 65 watts power draw.

Assuming I'm right, vs. AMD's RX 460, GP107 would have both a larger deficit in DX12 and larger advantage in DX11 games (when compared to RX 480 vs. GTX 1060) based on the few reviews I've read, making the choice between the two cards all the more difficult.

Yeah I forgot that Nvidia has ROP count tied to memory controllers.

Either way though the GP107 should be somewhere between 50% and 66% of a 1060 (assuming equal clocks), it all depends upon whether or not the 1060 is ROP or bandwidth limited.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |