Even with Pascal just around the corner, Nvidia isn’t letting up with new graphics card launches. After AiB partners launched a series of low power GTX 950s based off of the GM206, Nvidia has launched their own card. Slotting into the professional lineup, the new Quadro M2000 features a full GM206 and is pretty much a 75W GTX 950 rebranded for the professional use.
At 768 CUDA cores and 1.3 TFLOPS of SP performance, the M2000 is the lowest member of the Maxwell Quadro lineup. The card replaces the Kepler-based K2200 which featured 640 CUDA cores. The card features 4GB of GDDR5 over a 128bit, offering 106 GB/s of bandwidth, 26GB/s more than it’s predecessor. Just like the recent slew of 75W GTX 950 cards, the M2000 won’t require a PCIe power connector, coming in with a slim single slot form factor with four DisplayPort 1.2.
With Maxwell just over a month away, it’s a really odd time for Nvidia to be releasing a new card. This leads me to suggest that Nvidia may not have any low-end Pascal cards out for the rest of 2016 except maybe a few mobile chips. From what we’ve heard, only GP104 will be unveiled at Computex so GP106 may not arrive for a while. A single Quadro M2000 will set you back 569 EUR.
Even with pascal so close, Nvidia and their partners are still pushing out new Maxwell based cards. Over the course of the last month or so, ASUS, MSI, and EVGA have all come out with new 75W TDP GTX 950 graphics cards. These new cards delete the PCIe 6 pin power connector, making them a more flexible option. Today, Gigabyte has quietly joined the party with their own 75W GTX 950 offering.
The GIGABYTE GV-N950D5-2GD graphics card is based on the same old GTX 950 GM206 core with 768 stream processors, 48 texture units, 32 ROPs, and a 128-bit GDDR5 memory interface. Somehow, Nvidia has been able to reduce the TDP from 90W to 75W without reducing base performance. Unlike the versions with power, the new card features a tamer factory overclock, with only 1051 Mhz base and 1228 MHz boost clocks respectively. The cooler is a relatively standard slot with 1 fan. Without the need for a PCIe power connector, it allows the card to be used to upgrade systems that feature a weaker PSU. The value of such a card is diminished though with Pascal being so close. In my view this suggests that Nvidia may leave the low end untouched for a while, with only GP104 launching anytime soon. This may leave AMD with the Polaris 11 relatively unchallenged, only doing battle with older Maxwell cards.
All throughout this week, we’ve been treated to leak after leak and release after release of new Polaris and Pascal information. After we’ve gotten die shots of GP100 and GP104, what may be another Pascal chip has surfaced. According to Videocardz, the Drive PX 2 module Jen-Hsun Huang waved in front of our faces might actually be Pascal this time around. More specifically, it might be GP106 or GP107.
Nvidia took some heat earlier in the year when CEO Jen-Hsun Huang took to the stage and waved a Drive PX 2 around that was Maxwell. This time around though, it looks like Pascal has come far enough along to make it into engineering samples. Based off of size comparisons to GM206, the Drive PX 2 GPU is too small in comparison. This means it is either the budget GP106 or the smaller mobile GP107.
This stems from the fact that the die is really similar to GM206 and given what we know of GP104 and GP100, the GP106 and GP107 should be of a similar size to GM207 and GM206. I believe that given the MXM format, we may be seeing the GP107 as that chip is meant for the mobile market. Using a chip that hits both the lucrative laptop and car markets would make a lot of sense. Either way, budget Pascal may be coming sooner than expected.
Even with Pascal just around the corner, Nvidia and their partners are continuing to release some new Maxwell GPUs. Today, we’ve been treated to a new lineup of Nvidia’s GTX 950 from EVGA. The biggest and one of the few differences between the new GTX 950’s and the old ones are the reduction of TDP from 90W to 75W. In all other regards, the cards are just the same old GTX 950 we’ve come to expect.
Like the ASUS and MSI cards from before, these new GTX 950’s are somehow able to reduce their power consumption from 90W down to 75W while using the same die. Some of the 75W TDP models retain the 6pin PCIe power connector, allowing them to reach 1127Mhz base and 1317Mhz boost clock. For the ones that delete the 6pin PCIe power connector, they peak at a lower 1075Mhz base and 1253 boost.
For the segment of the market the GTX 950 targets, any reduction of power consumption is important. Not everyone has a PSU with a 6pin connector and removing it makes sense if you can keep the same performance characteristics. There is a total of 8 new cards EVGA has launched today and they are all available now.
For graphics cards, 75W is a golden number as it dispenses with the need to have a separate power. This allows users to avoid a PSU upgrade and broadens the market for the card. Originally launched at 90W, it looks like Nvidia has managed to trim an additional 15W savings to produce 75W GTX 950s. First started off by ASUS, MSI is getting into the 75W GTX 950 as well with 2 new additions to their lineup.
Dubbed the GTX 950 2GD5 OCV2 and GTX 950 2GD5T OCV3, they will replace/supplement the 90W GTX 950 2GD5 OCV1 and GTX 950 2GD5T OCV2 respectively. Both cards are based on NVIDIA’s GM206-251 GPUs, with the 251 indicating either a special bin or a new process that Nvidia is using. Of course, the chips are still the same GTX 950 with 768 CUDA cores, 48 TMUs, and 32 ROPs with 128-bit GDDR5 memory interface fed by 6.6Gbps of 2GB GDDR5. Both cards are factory overclocked to 1076Mhz stock and 1253Mhz Boost.
Between the OCV2 and OCV3, the only difference is in the cooler and form factor. The OCV2 uses a single fan and is geared towards mITX of other small form factors. The OCV3 sports a larger dual-fan cooler and is longer to boot. Both cards feature hardware-accelerated decoding and encoding, making them good choices for HTPC or to upgrade an existing desktop system for some moderate gaming. No word yet has been revealed about pricing but expect it to fall near MSI’s current offerings.
Maxwell has been Nvidia’s most efficient GPU architecture to date, offering great performance/watt. While the lower end cards aren’t quite as power efficient as some of their bigger siblings, the GTX 950 is as good as it gets, offering decent 1080p performance at 90W. That, however, is set to change as Nvidia is preparing a 75W GTX 950, one that ASUS is launching first.
Due to the fact that it only requires 75W, the ASUS 950 can afford to draw all of its power straight from the PCIe 3.0 x16 slot, cutting down on cable clutter and PSU requirements. Despite losing out on power headroom, ASUS has still managed to give the card a factory OC mode of 1051 MHz Base and 1228 MHz Boost A more tame default Gaming mode has it at 1026 MHz Base and 1190 MHz Boost. This all the while keeping the same GM206 core of 768 CUDA cores, 64 texture mapping units, 32 ROPs and a standard 2GB of GDDR5 VRAM.While the boost clocks are pretty good, it’s important to keep in mind that those are the max figures and with the cut-down power delivery, it’s unlikely the card will keep or even hit those numbers, especially under load. As with all silicon, Nvidia and ASUS is either binning chips heavily, 28nm is getting much better (unlikely this late into the node) or simply cutting enough voltage while maintaining base clock speeds.
This late into the generation, it looks like Nvidia is prepping for an efficiency battle with AMD’s upcoming Polaris which set to be many times more efficient than current GCN cards. The new GTX 950 may have to do battle on it’s own for a while till Nvidia get their own competing Pascal cards out.
Originally launching in both a 2GB and 4GB variant, Nvidia is reportedly planning to discontinue the lower capacity model. By offering only a 4GB tier, Nvidia is hoping to make the card more attractive to buyers as they will only see the 4GB version. At this point in time, there is no word yet if the 4GB 960 will keep its current price or drop to fill in the void left by the departing 2GB model.
The GTX 960 features the full GM206, Nvidia’s budget Maxwell die. While the card does decent against AMD’s R9 380, it does fall behind a bit in terms of overall performance. With the launch of the GTX 950 as well, the 960 has become even more of a niche product. The 950 features only 256 fewer shaders and 12 TMUs, not a large margin by any means, placing its performance to near 960 levels. With such competition, it is understandable why Nvidia will try to differentiate the card more by only having a 4GB model.
The biggest question is whether or not the GTX 960 will actually need 4GB of VRAM. While 4GB might be needed for 1440p, the 960 is solidly a 1080p performing card. That has historically been the domain of 2GB of cards and by the time 4GB is required for 1080p, the GPU core of the 960 may well be lacking. One also must consider the fact the 950 also has a 4GB model and would age about the same as the 960. Both cards are also limited by the 128bit memory interface which may hinder the use of such a large frame buffer.
Undoubtedly though, the extra frame buffer would make the 960 more future proof if only just. It will be interesting to see if Nvidia does follow through with this move in the end. We will follow this story as it develops and bring you more information as it arrives so stay tuned!
Thank you HWBattle for providing us with this information
While various reports and rumors have suggested that the GTX 950 is well on its way, additional details have come out about the card and it’s GTX 950 Ti older sibling. First of all, the report suggests that the GTX 950, which we had long known would arrive in late August, is scheduled for a launch on the 20th, a little under a week away.
Bigger news though is the confirmation of the GTX 950 Ti. According to the source, Nvidia is still developing the GeForce GTX 950 Ti. This card will be based on the GM206 and slot in between the GTX 950 and GTX 960 given its name and might receive a 4GB version. Given that the GTX 950 has 768 shaders and the GTX 960 has 1024 shaders, the 950 Ti is pretty much guaranteed to have 896 shaders. This means Nvidia will have 3 cards packed tightly together in terms of performance. TMU count is expected to be the same between are 3 GM206 siblings but ROP counts for the 950 Ti are still unknown.
With 3 cards so tightly packed together, Nvidia risks confusing consumers and can only really differentiate on price. The shader deficiency between the 980 and 970 hasn’t kept the 970 from nearly matching the 980 and the smaller differences between the GM206 cards probably means even smaller performance disparity. The cut down 950 and 950 Ti will allow Nvidia to get rid of imperfect GM206 dies though. While more budget cards are welcome, Nvidia would probably be better served with a card to slot in the large gap between the 960 and 970, served only by AMD.
With Nvidia’s GTX 950 expected later this month, it’s not surprising to see that some listings and pictures of the card have started to emerge. Today we bring you the first listing for a GTX950, the ASUS STRIX GTX 950 DC2OC. As expected from the model name, the card should be quite similar to the GTX 960 Strix with an oversize DirectCUII cooler on a small PCB. There are multiple listings out now at about €240+, but we can expect the launch price to be closer to €150 to replace the 750Ti as the currently listed prices surpass those of the 960.
More interesting is the first glimpse of the GTX 950 from PNY. From the packaging, we get the usual 3 years PNY warranty along with “High-Resolution Graphics Cards” which probably means it can support 4K output. The cooler is relatively simple and we can see that the PCB is larger than that of the 750Ti and closer to the 960 though we don’t know if it’s a custom PCB. The low power nature of the 950 does betray itself as we can see that the cooler is a simple radial aluminum heatsink, likely without any heatpipes. Given that the 960 has a much beefier heatsink, the extra PCB is probably there to support more power delivery components and support the larger heatsink.
The GTX 950 is based on the GM206 with 768 Maxwell shaders on a 128bit bus with 64 TMUs and likely 32 ROPs. With performance likely to fall around the R7 370, we can expect pricing to be about the same as well. Nvidia has already cut the price on the 750Ti so the 950 should be arriving imminently.
Thank you Videocardz for providing us with this information
We brought you news last week that Nvidia was preparing a more budget friendly cards in the form of the GeForce GTX 950 and 950Ti. Now, more information has been revealed about the upcoming cards, pointing to a singular release for the GTX 950 alone. As expected the GPU will be the GM206-250, supposedly a cut-down version of the GM 206-300 used in the mainstream GTX 960.
No information about the core configuration has been revealed yet, but we can make some educated speculation. As replacements for the Maxwell 750/750Ti, it is unlikely the 950 will be the same as the 750 which has a core configuration of 512:32:16 (Shaders:TMU:ROP:). That either points to 640:40:16 like the 750Ti or something with more oomph like 768:40:24. Remember, if Nvidia does plan on a 950Ti in the future, it has to leave enough room for it to make sense, so the 950 launch may be a good indicator to see if we ill get a Ti version in the future. The GTX 960 comes in at 1024:64:32 so at the very least, the 950 likely won’t hit higher than 896:40:24 in order to get the most out of imperfect GM206s. On the other hand, purposefully lasering off parts, especially on the 28m process where yields are high is also suboptimal.
On the memory side of things, the information suggests that the 950 will come in either a 2GB or 4GB configuration, much like the 960 it is based on. That would put memory capacity at double the 750/750Ti and the previous 650 series. The memory numbers suggest a 128Bit bus, and it is unlikely it will be any higher as the 960 only comes with a 128Bit bus. The 2GB-4GB frame buffer points to a target resolution of 900 or 1080p like the 960 but likely at a lower playable quality. Final specifications are DVI, HDMI and 3 DisplayPorts, similar to the 960. With no launch date yet, it will be interesting to see what Nvidia brings to the table as the AMD side of things with the R7 370 is less than optimal.
Thank you Videocardz for providing us with this information
VideoCardzhave managed to get a hold of the new Nvidia GM107 GPU die layout. The GM107 is based on the next-generation Maxwell GPU architecture and according to rumours it contains a very different layout to its predecessor “Kepler”. In the GM107 die, which is expected to form the GTX 750 and GTX 750Ti, there are five SMM (streaming multiprocessors) blocks which each contain 128 CUDA cores. This makes it different to Kepler where one SMX unit contained 192 CUDA cores. If these rumours and the above image is true it means that the GM107 features 640 CUDA cores. Apparently all five SMM units are enabled on the GTX 750 Ti meaning it gets all 640 CUDA cores while the GTX 750 is rumoured to have one of these disabled meaning 512 CUDA cores in total.
Apparently the GM107 and another smaller (read: slower performing) chip, the GM108, are both to be built on the 28nm process with the Maxwell architecture. In terms of the higher end the GM206, GM204 and GM200 are the three GPU dies that will succeed the GK106, GK104 and GK110 respectively. These GPUs are expected to be built from Maxwell with a 20nm, not 28nm, process.