Search the Community
Showing results for tags 'soc'.
-
A new rumor mill indicates Intel is hard pitching its offerings to Microsoft for the next Xbox. If the rumor has some leg, this will be the first time in 20 years the chip veteran could find its way in a gaming console. Quotes Considering the mail source of the rumor (see below), don't put too much hope in it -- it's still hearsay with generous amount of assumptions. On the other hand, a potential deal for Intel would reinvigorate their further GPU R&D that would also be a positive contribution to the PC market competitive landscape. Sources: https://www.techpowerup.com/318820/intel-looking-to-grab-microsoft-xbox-semi-custom-soc-business-from-amd https://www.tweaktown.com/news/95974/intel-wants-to-be-inside-microsofts-next-gen-xbox-console-would-built-in-the-usa/index.html
- 49 replies
-
- intel inside
- xbox
-
(and 1 more)
Tagged with:
-
Alright, before I go and daily these settings, could anyone tell me if they're safe? PBO is configured with Auto Scalar & Default Limits Asus Core Performance Boost is enabled. Curve Optimiser set to -15 on almost all cores and -10 on two of them (this is what I found to be stable, -30 wouldn't even boot). Core, SOC Voltages are set to Auto, reading as 1.353V and 1.080V in CMOS respectively. In single-threaded workloads, VID reads between 1.2V and 1.26V Clocks range between 4.69Ghz (nice) and 4.95Ghz in single-threaded workloads, steady at 4.6Ghz all-core in Cinebench Max temperatures are around 87⁰C (it's a bit toasty but I think this is all my NH-D15 can do at over 210W all-core) DOCP at 3600Mhz, Auto Voltage reading at 1.352V System Specs: AMD Ryzen 5950X (PBO Enabled, Core Enhancements Enabled) Noctua NH-D15 chromax Corsair Vengence LPX 3600Mhz/C16 (DOCP Enabled) Asus Crosshair Dark Hero VIII X570 Samsung 980 Pro 1TB x2 LG Optical Drive Hotplug 2.5 inch SATA Bay Samsung 870 Evo 2TB Nvidia RTX 3090 Founders LSi 9211-8i HBA 5x Seagate IronWolf Pro 6TB Corsair HXi 1000W PSU
-
Summary MediaTek rumored to be partnering up with NVIDIA for developing low power GPU for their mobile SoC in early 2024 Quotes My thoughts I don't have much to say about the NVIDIA x MediaTek news. I kinda don't think this is a big surprise, in fact they did that RTX on ARM demo with a MediaTek chip back in 2021 which was during the NVIDIA ARM acquisition drama, but at the same time I do hope this is not NVIDIA trying to find a loophole to strangle ARM chips market silently and I hope this doesn't spark any gatekeeping problem for basically everyone else that are partnering up with MediaTek for their mobile chip or anything else, especially that MediaTek still have some partnership with AMD in some spaces like Network solution for AMD mobile platforms. But we shall see whether this is going to be actually happening or not Sources - https://www.digitimes.com/news/a20230515PD212/gpu-mediatek-nvidia.html - https://www.tomshardware.com/news/mediatek-to-adopt-nvidia-gpu-tech-report - https://9to5google.com/2023/05/16/mediatek-nvidia-gpu-report/
-
I have XMP enabled, also I have PBO on as well and I hit a barrier with 3200MHz on 1.15V SoC when i undervolt it to 1.1V (default) i can get to 3533MHz, but it gets unstable and past that point i can not go further, unless i now for real undervolt it a bit lower so i can get to 3733MHz. 1.05V on SoC seems to be too low -it gets me in but crashes real fast but it also worked on 1.075V but again unstable. What would be the sweet spot with something as Patriot Steel Viper 3733MHz RAM to work stable on at least 3400-3600MHz? They are Hynix C-die rams 2 moduls VDDG was set between 950mV and 1050mV cLDO VDDP was set to around 1050mV and DRAM Voltage to 1.35V to ensure stability. But with that said, SoC Voltage was the crucial thing i was playing around with, with PBO enabled (i think) doesnt like higher voltage or it wont even post with freq beyond 3200MHz with 1.15V Basically i could increase higher SoC voltage and as long as it doesnt go past 3200MHz it will be alright, but beyond that i'll have to lower it.
-
Arm CPUs are taking over. Apple Silicon showed us that desktop computers need not be power hogs - Why haven't AMD, Intel, and Nvidia done the same, and would you want it?
-
From the album: Graphics Card GPU
GV-N560SO-1GI-950 REV : 1.0 -
Source: http://www.anandtech.com/show/11201/qualcomm-snapdragon-835-performance-preview/6 It seems that recently Qualcomm invited many reviewers to its San Diego HQ to test phones built around its Snapdragon 835 SoC. These are development phones made by Qualcomm for hardware and software testing, so don't get too excited. The new Snapdragon 835 is an octa core big.LITTLE CPU based on Samsung's new 10 nm LPE lithography, with the power cluster containing 4 Kryo 280 cores at 2.45 ghz and the efficiency core containing 4 Kryo 280 cores at 1.90 ghz. These cores aren't completely new designs, as evidenced by the shared "Kryo" name, but they have some improvements that should bring about higher IPC. It also has a new GPU, the Adreno 540, which should bring a pretty substantial performance improvement compared to the Adreno 530 from the Snapdragon 820 and 821. Anandtech was able to perform quite a few tests. Naturally, I'm not gonna put everything here, but I'll put a few of the benchmarks in order to give you a good idea of how the 835 performs. First up is Geekbench 4's single threaded integer performance. Geekbench 4 - Integer Performance Single Threaded Snapdragon 835 Snapdragon 821 (% Advantage) Snapdragon 810 (% Advantage) AES 905.40 MB/s 559.10 MB/s (61.9%) 714.47 MB/s (26.7%) LZMA 3.13 MB/s 2.20 MB/s (42.3%) 1.92 MB/s (63.0%) JPEG 16.80 Mpixels/s 21.60 Mpixels/s (-22.2%) 12.27 Mpixels/s (36.9%) Canny 23.60 Mpixels/s 30.27 Mpixels/s (-22.0%) 23.63 Mpixels/s (-0.1%) Lua 1.84 MB/s 1.47 MB/s (25.2%) 1.20 MB/s (53.3%) Dijkstra 1.73 MTE/s 1.39 MTE/s (24.5%) 0.91 MTE/s (90.1%) SQLite 53.00 Krows/s 36.67 Krows/s (44.5%) 33.30 Krows/s (59.2%) HTML5 Parse 8.67 MB/s 7.61 MB/s (13.9%) 6.38 MB/s (35.9%) HTML5 DOM 2.26 Melems/s 0.37 Melems/s (510.8%) 1.26 Melems/s (79.4%) Histogram Equalization 52.90 Mpixels/s 51.17 Mpixels/s (3.4%) 53.60 Mpixels/s (-1.3%) PDF Rendering 50.90 Mpixels/s 52.97 Mpixels/s (-3.9%) 43.70 Mpixels/s (16.5%) LLVM 196.80 functions/s 113.53 functions/s (73.3%) 108.87 functions/s (80.8%) Camera 5.71 images/s 7.19 images/s (-20.6%) 4.69 images/s (21.7%) It's quite apparent integer performance was something Qualcomm focused on for Kryo 280. While the improvement over Snapdragon 820 varies from test to test, it's quite apparent Qualcomm definitely has improved. Kryo's integer performance was relatively weak compared to other SoCs, so it's nice to see that Qualcomm has caught up with the rest of the market here. Moving on, we have Geekbench 4's Floating Point Performance. Geekbench 4 - Floating Point Performance Single Threaded Snapdragon 835 Snapdragon 821 (% Advantage) Snapdragon 810 (% Advantage) SGEMM 11.5 GFLOPS 12.2 GFLOPS (-5.7%) 11.0 GFLOPS (4.2%) SFFT 2.9 GFLOPS 3.2 GFLOPS (-9.7%) 2.3 GFLOPS (25.2%) N-Body Physics 879.6 Kpairs/s 1156.7 Kpairs/s (-24.0%) 580.2 Kpairs/s (51.6%) Rigid Body Physics 6181.7 FPS 7171.3 FPS (-13.8%) 4183.4 FPS (47.8%) Ray Tracing 232.6 Kpixels/s 298.7 Kpixels/s (-22.0%) 130.1 Kpixels/s (78.7%) HDR 7.8 Mpixels/s 10.8 Mpixels/s (-27.6%) 6.4 Mpixels/s (21.9%) Gaussian Blur 23.4 Mpixels/s 48.5 Mpixels/s (-51.8%) 21.9 Mpixels/s (6.7%) Speech Recognition 13.9 Words/s 10.9 Words/s (27.5%) 8.1 Words/s (71.4%) Face Detection 513.8 Ksubs/s 685.0 Ksubs/s (-25.0%) 404.4 Ksubs/s (27.0%) Strangely enough, floating point performance seems to have regressed relative to Kryo. This shouldn't be too bad, since Kryo had very strong Floating Point performance, but it still is definitely a strange thing to see. I'll skip the PCMark and Java Script tests, but the 835 is essentially near if not at the top in PCMark and is consistently number 1 or 2 out of the non-Apple SoCs in the JavaScript tests. Now for the GPU tests. I won't bore you guys with all the test. Here's one that's relatively representative of the other tests: As you can see, the new Adreno 540 GPU is now probably the most powerful smartphone GPU, beating out last year's Adreno 530 as well as the PowerVR GPU used in the iPhone 7 and 7 Plus. Now, onto something a bit more interesting... Power consumption! Qualcomm Power Testing - Device Level w/Fixed Workload Power Consumption SD820 Reference Phone 4.60W SD835 Reference Phone 3.56W The Snapdragon 810 really failed hard at power consumption. More insight was recently provided in Anandtech's Kirin 960 review, which you can find here. It showed, the 810 consuming around double the power of other SoCs at the same number of cores utilized (It was consuming 8 watts of power at 3 cores utilized). Snapdragon 820 was a huge jump from this, and it seems Qualcomm has further improved power efficiency with the 835. This is definitely a good step forward. The Snapdragon 835 is looking like a great evolutionary upgrade to the 820. I look forward to seeing the next generation of smartphones utilizing this SoC.
- 42 replies
-
- qualcomm
- snapdragon
-
(and 2 more)
Tagged with:
-
Can someone suggest me the best cpu under $30....LGA 775 socket type is what i want.
-
ARM slides detailing the future of their IP has been leaked. ARM has long had a fairly slow release cadence but it seems they've picked up speed with them announcing A75, A55 and G72 (high performance CPU, high efficiency CPU and high performance GPU respectively. These will be successors to A72/A73, A53 and G71. A53 has seen wide-spread use by multiple semi-conductor companies and has been employed in numerous designs. The A53 was actually announced in 2012, so as you can imagine it was long overdue for a successor. With up to 2x higher performance, 15% more efficient and being more configurable, the A55 will also be used in numerous designs from entry level to flagship as efficiency clusters. The A75 will be a welcome replacement for the A72/A73. A73 was a good efficiency gain over A72 but also regressed in performance in some areas. It will offer 20% more performance than A73 with the same efficiency The G71 marked a huge performance improvement in ARM's graphics endeavors with scalable designs that finally matched what Qualcomm's Adreno chips could offer where we saw Samsung implement an MP20 design in their Exynos 8895 that actually exceeded Qualcomm's Snapdragon 835 (which has their latest Adreno 540) in graphics performance. The G72 will continue this trend with 25% better efficiency and 20% better performance density. Lastly, they'll also reveal more about their new DynamIQ cluster designs which is an improvement over conventional big.LITTLE (the use of multiple clusters of varying design or performance characteristics) which has been the paradigm in ARM designs and therefore the de facto standard in Android devices. All this will be revealed at the ARM Tech Day on Monday the 29th. Source: https://www.nextpowerup.com/news/36256/leaked-slides-detail-arms-new-cpu-and-gpu-designs/ -------------------------------------------------------------- UPDATE: So ARM had its Tech Day and the embargo has been lifted. This means Anandtech has done an analysis of what we can expect by dissecting the PR numbers. It seems the performance of the new A55 cores are hitting a ceiling by remaining In-Order but the cost for going Out-of-Order in power consumption might be too great for now. DynamIQ is a huge shift in how ARM does multicore processors. We've been used to a maximum of 4 cores per cluster and using 2 clusters with each cluster having the same cores. With DynamIQ things will change. It will provide the ultimate flexibility with clusters with up to 8 cores and you can now mix your cores however you want meaning a 7 x A55 and 1 x A75 is possible now, for example. This change also improves L2 cache latency by 50% simply by moving the L2 cache closer to the cores. Anandtech's analysis, as always, is worth a read. There is sadly not anything on the new G72 GPU though. Source: http://www.anandtech.com/show/11441/dynamiq-and-arms-new-cpus-cortex-a75-a55
-
Google may been working to design custom SOC's for future mobile devices. A recent hire at Google indicates big changes are coming for future versions of the Pixel phone. It seems that google may have just hired the ex-apple micro-architect, Manu Gulati, who worked on the development of apples SOC's for nearly eight years. this is a completely new path for google as they what do you guys think? will google be able to utilize the knowledge of an ex-apple employee and leap-frog apple in terms of performance? Where do you think google will start, in the creation of SOC's? Sources: https://www.theverge.com/2017/6/13/15791918/google-hires-apple-chip-architect-pixel-phones http://wccftech.com/pixel-custom-socs-google-hiring/
-
Source: http://www.anandtech.com/show/11569/imagination-technologies-formally-puts-itself-up-for-sale As some of you may remember, close to 3 months ago Imagination, the company that has been designing the GPUs Apple has been using in their SoCs for generations now, announced that Apple had informed them that in a year or two Apple would stop using Imagination PowerVR GPUs in its SoCs. Due to around half of Imagination's revenue coming solely from Apple, this was a huge bombshell and shortly after Imagination stock fell A LOT. Before the announcement imagination stock was priced at around $280 per share, but immediately after it dropped to $100, though now it's managed to climb back to about $140 per share, or around half of what the stock was priced at before the announcement. Around a month after the initial announcement Imagination announced that they would be selling off the other parts of its business, focusing solely on its most profitable business, the GPU business. However, now it seems that Imagination has instead decided to, instead of trying to survive off its GPUs, put itself on sale. What would be quite interesting is if, in the end, Apple would turn out to be the one who purchases Imagination. A lot of features from the PowerVR architecture are used in iOS App Development, including PowerVR Texture Compression, so moving away from PowerVR wouldn't be an easy feat for Apple. However, with Imagination's market cap currently at around $500 million, or about half of what they were at before the initial announcement, if Apple bought them now they could purchase the company for significantly cheaper than what it might have initially cost them. At any rate, I'm quite interested to see who might end up purchasing Imagination. There are lots of companies out there that might be potential buyers and regardless of who will end up purchasing Imagination this might end up changing the mobile SoC market in the future.
- 47 replies
-
- imagination
- apple
- (and 4 more)
-
Which SOC is superior and why? They seem pretty equal on paper
-
So this kind of flew under the radar but tesla has unveiled it's complete compute hardware solution today on a livestream. Quite interesting. Some specs I'm still watching it so I might add more later. The previous computer was a custom Drive PX2 from Nvidia with the following specs: 2x Tegra X2 ( 4x Denver 8x Cortex A57) 1x Parker GPGPU 4 FP32 TFLOPS 10-12 DL TOPS (Compared to 72 TOPS on the new chip) 60w TDP I should point out that Nvidia's announced but unseen (?) Drive PX Pegasus is spec'd at 320 TOPS vs the 144 total TOPS of this system but Pegasus is at a TDP of 500W vs TDP of 100W here. CPU vs GPU VS Neural Network processing power What the neural network is doing How chip TDP's are managed at the transistor level (picojoules) Actual power dissipation at current usage levels (usage is low right now, the chip has tons of capacity) Benchmarked using camera data They also say cost is 80% of Nvidia's solution (semi-custom Drive PX2) Timelapse footage of new hardware in action: Servers each with 4 FSD Computers used for training validation THE MASTER PLAN He's not joking. No seriously
-
Apple’s finally done it – They’re moving away from Intel and for the first time going it entirely alone with an ARM-based processor. Can they make it work? HOW can they make it work…?
-
Original articles: https://wccftech.com/amd-zen-vega-semi-custom-soc-powers-gaming-consoles/ https://www.pcgamer.com/amd-built-a-powerful-semi-custom-ryzen-vega-soc-for-game-consoles-and-pcs/ AMD's announcement: https://community.amd.com/community/gaming/blog/2018/08/03/new-amd-semi-custom-soc-combines-the-power-of-amd-ryzen-cpu-and-amd-vega-gpu-for-gamers-in-china AMD has released a new custom SOC (system on a chip) for the China market primarily aimed at home consoles. While current APUs on the general market (2200G, 2400G, PRO 2400GE) have Vega 8, 11, and 11 respectively, this new SOC comes with Vega 24. This new SOC was recently revealed at the China Joy event. Specs: Base frequency: 3.0 GHz, quad-core, 8 thread 24 Vega CUs with up to 8 GB GDDR5 attached, GDDR5 memory controller on-chip L2 Cache: 2 MB L3 Cache: 8 MB 4 TFLOPS at max clock " Currently, AMD is working with Zhongshan Subor to create a new gaming console for the Chinese market which was displayed at ChinaJoy 2018. They will first introduce a gaming PC with the SOC in late August, followed by the gaming console release in late 2018. Pricing of the PC is expected at 4998 Chinese Yuan which equals around $600-$700 US. " My personal speculation is that this is a good trial run for what we'll possibly see in the next generation of Microsoft and Sony console, at least as a launch platform. AMD's recent forays can only mean good things for the company in terms of its bottom line. Content Edit 1: If we're comparing Vega cores, the 2400G's 11 CUs are clocked at 1.25GHz while the SOC's 24 CUs are clocked at 1.3GHz according to the article. Content Edit 2: Added link to AMD's official blog post concerning the SOC. Content Edit 3: Clarification of memory controller.
-
Hi guys bassicly i been wondering is there any news about soc that will come after kirin 659. Since there is been like no improvement for huaweis mid range phones since kirin 650 Iam expecting some kind of actual improvement this time.
-
As I've said before: Intel is not built of fools. Intel wants more business, and as Kirzanich says here, if it wins Intel the business, it's willing to make ARM cores for as long as needed (prefaced by saying Intel's average product lifetime is 12 years). Prediction: in 5 years, the mobile world will still run ARM, but Intel will be a player to be reckoned with at last. Sorry if this seems like this should have been combined with my FPGA post, but I feel these are two very big pieces of news that deserve their own discussions. "We have no plans, none! I don't care what you hear, what the competition says, or anything else. There's no plan to yank ARM out and stick in IA and make you guys change all your programming models, absolutely none. I do believe there will be some additional products in the future, and I want you guys to take advantage... Virtualization, multiple operating systems on an FPGA per core, ... "
-
Does anyone knows of a good comprehensive guide to embedded and low end intel CPU´s ? (and even amd) I was looking at devices like this Voyo machine http://www.gearbest.com/tv-box-mini-pc/pp_292915.html This machine has its flaws, the Wireless is 2.4Ghz only, the windows version it comes with is activated for china only so the key is only valid if you install a chinese windows 10. and some have reported heat issues, But the thing is to mesure the actual capacity of the CPU, its hard to find these machines with a top of the line Atom and decent ram and storage, most of them are the usual 2gb ram 32gb storage (which you get 12-16 after a full install and set up), so the fact that its using the X7 Z8700 and has 4gb of ram is interesting, But you can get a N3050 NUC and some ram to put in it for the same cash or a stripped down beebox from asrock with a N3150 that just needs ram and a HD The thing is that i know Intel has several low end chips for HTPC´s, NAS, Embedded and tablet use, The J class celerons the X3/5/7 Atoms and the N class celerons. The problems is that i cant seem to find a clear comparison between those chips and desktop counterparts and most importantly between themselfs. The only thing i know is that some of the higher end X7 and N series celerons can actually stand up to a older Core 2 duo and in some cases the X7 can actually surpass some mobile I3´s and can even move dolphin (Gamecube emulator) at a decent framerate. But theres nothing clear there. Anyone knows somewhere i could research this ? (and by research i mean im lazy and want a clear comparison )
-
I just noticed an ad for the UDOO X86 Kickstarter page. They claim this $89 board can drive 3, 4K displays simultaneously with an Intel quad core at 2 - 2.24GHz, and 4 - 8 GB of ram. Not sure how much I trust that claim. From the description it sounds like an Atom x5 or x7. Part of me would like to see Linus back the Kickstarter so we can see how the final product turns out when/if it comes out. What are your thoughts?
- 5 replies
-
- udoo x86
- kickstarter
-
(and 3 more)
Tagged with:
-
This thing looks neat I wonder if it is any good! http://www.amazon.com/VIA-Embedded-EPIA-M920-12Q-Mini-ITX-Processor/dp/B00D5KC5MS
-
(Before you read about the pricing and the availability...I'm from Hungary, so prices may vary) I often get asked to suggest a budget computer build for home use, and I usually suggest an AMD A4 based build (Usually the cheapest one available). When you read that you're like: OMG! One of the worst available processors right now... But it's actually not that bad. 2 cores/threats and an integrated graphics processor that's able to play 1080p. From the feedback I've got it's more than enough for home use. Where I live...You can buy an FM2+ A88X Asus Mobo for about 55$ and pair it with an AMD A4-4000 that costs 30$. If you put it together that's only about 85$. However I'm ALWAYS tempted when I see Integrated CPU motherboards... Such as the Asrock J1900M motherboard, with a 4 core Intel Celeron CPU clocked at 2GHZ onboard...For only around 65$. I know we lose CPU/APU upgradability with the Embedded CPU motherboard, but usually the buyer of such budget computer consider 4 cores a bit overkill already, seeing how they were fine with their E4300 s775 chips for so long. So, what do you say? Was I right suggesting the FM2+ socket mobo with an FM2 APU? Or maybe I should have told them to get that Asrock one?
- 9 replies
-
- integrated
- motherboard
-
(and 4 more)
Tagged with:
-
From Altera Not to familiar with this FPGA SOC stuff so don't know what to comment on. Still, it's nice to see technology going forward. http://wccftech.com/intel-altera-offering-hbm2-fpga-sip-stratix-10/
-
I've heard a rumor that some AM1 boards support registered ECC. I've also read statements that say it only supports unregistered and one article that says am1 is incapable of ECC even if the processor is. I've been unable to confirm anything with a secondary source so I was hoping someone here definitively knew one way or the other. I've also been looking at avoton and atom SOC options and I'm generally looking for relatively small and cheap options that might support registered ECC. I understand the details and that I can function just as well without it, and if it comes down to it and there are no such options that's likely what I'll do. I do however happen to have 32gb just laying around and I thought I might put it to good use. my original budget was 650$ before drives and I'm hoping to bring that under 400$ before drives. I was thinking maybe a 400w fanless seasonic platinum psu at first now I'm thinking that's overkill at the price point so I'm open to suggestions and alternatives. I know that I can probably build a super cheap am1 at around 150$ (contingent on ECC) but while I'd like to keep this cheap as possible I'd also like it to be in a case. I estimate I'll be using 6 x 5tb drives, 2 of which would be for backup, and I probably have a SAS card or raid card in with my spare parts if I need it. I don't expect any build guides or logs, I can do the leg work and due dilligence on everything, though I hope the extra details pre-emptively answer expected questions. I really just need clarification on cheap motherboards that support ECC if any, or if I'd save money avoiding it considering the price of compatability. edit: Wanted to add purpose is mostly for transcoding, non-time sensitive and not particularly data sensitive but who ever wants to lose data right?
-
Qualcomm Announces Three new SoCs and a Gigabit LTE Modem. The three SoCs are, Qualcomm Snapdragon 635 Qualcomm Snapdragon 625 Qualcomm Snapdragon 435 Snapdragon 625 The Snapdragon 625 is the first mid-range chip from Qualcomm to be built on a 14nm FinFET process, which can save up to 35 percent power consumption compared to the previous generation. It has an octa-core Cortex-A53 CPU, and it comes with an integrated X9 LTE modem with upload speeds of up to 150 Mbps. It also includes an Adreno 506 GPU, which supports the brand-new Vulkan API. The chip has support for dual 24MP cameras on the back and 13MP cameras on the front. Snapdragon 435 The Snapdragon 435 is also an octa-core Cortex-A53 CPU that is matched with an Adreno 505 GPU, and it has support for dual 21MP cameras that can be paired with hybrid autofocus systems (phase detection and contrast auto-focus). An X8 LTE modem is integrated into the 435 SoC that brings download speeds of up to 300 Mbps and upload speeds of up to 100 Mbps. Snapdragon 425 Qualcomm’s Snapdragon 425 seems to be the direct successor to Snapdragon 410, which replaced the Snapdragon 400 in many low-end smartphones that cost around $100 or so. It comes with a quad-core Cortex-A53 CPU and an Adreno 308 GPU. It may have made sense to pick a generation-old GPU for the Snapdragon 425 SoC, due to cost constraints, but it’s a little odd that Qualcomm went with a two-generation old GPU here. Previous chips in this series used one-generation old GPUs. The chip comes with an X6 LTE modem that supports upload speeds up to 75 Mbps with 64-QAM and 2x10 MHz carrier aggregation. Snapdragon X16 LTE Gigabit Modem This new Cat. 16 modem is built on a 14nm FinFET process, which allows it to reach download speeds of up to 1 Gbps, although it only reaches upload speeds of up to 150 Mbps. 4G LTE has always been supposed to reach 1 Gbps speeds, just not from day one, so it’s not so surprising that we’re reaching this limit now. However, it’s still great to see it happen, and Qualcomm deserves credit for reaching this technical limit with LTE technology. Availability: The X16 gigabit modem and Qualcomm's three new SoCs are expected to arrive in devices in the second half of 2016. Source: Tom's Hardware.