During today’s GDC session on Epic’s Unreal Engine, NVIDIA CEO Jen-Hsun Huang dropped in as a special guest to announce NVIDIA’s next high performance video card, the GeForce GTX Titan X.
In order to capitalize on the large audien…
During today’s GDC session on Epic’s Unreal Engine, NVIDIA CEO Jen-Hsun Huang dropped in as a special guest to announce NVIDIA’s next high performance video card, the GeForce GTX Titan X.
In order to capitalize on the large audien…
One of the big announcements to come out of Mobile World Congress this week is regarding HTC’s new VR headset, called the Vive or ‘Re’ Vive, depending on where you see it, and their collaboration with Valve who appears to be providing the software backbone for the hardware. As part of our rush of MWC meetings, HTC invited me (only one person per publication) to experience the Vive developer kit with Steam VR. Having tested the Oculus Rift ‘Crescent Bay’ model at CES and having had a chance to sample Samsung’s Gear VR while here for a few minutes, it is interesting to feel how the VR experiences differ.
At this point in time, HTC is releasing next to no specifications about their developer kit, and the embargo placed today was purely for experiential testing. We were only able to take pictures of the device, and not the room in which it was being tested. Thus the following is purely from my perspective of using the kit and the pre-programmed demos that were provided. Overall each person was allowed 20 minutes to cycle through the demos, similar to that from Oculus at CES.
Functionally I was lead into a near empty room several meters square, put on the headset, and a HTC employee placed a set of headphones over my ears. When the demo started up, the guide held up two controllers, one for each hand. Within the headset, I saw the controllers represented on the screen and reached out to them. The headset and controllers both had a number of wires coming from them, but I was told that the end goal is to make the controllers wireless, and the headset would in the final version have only a single mini-HDMI cable to provide the bandwidth for the display.
The two controllers were designed such that one is for each particular hand with a depressed pad on each (kind of like the initial Steam box controller renders), along with a rear trigger and two grip buttons. Within the first demo, depressing the left trigger caused a balloon to appear from the controller, which I could then let go and punch away with either controller. After a few balloons the main demo began.
The first scene was a basic white field to describe the motions of the set up. I could move my head, walk about 2 meters in each direction before hitting an invisible wall (or an actual wall), as well as duck to get a different perspective from close to the ground. The white field was filled with columns moving up and down to get a sense of perspective.
The next scene was interactive – I was in a kitchen in front of a chopping board with vegetables, with a pot on the hob to my right and a robot helper ahead. With some mild instruction, I used the controllers and the triggers to pick up some of the objects. I put some bread into the pot and threw the rolling pin at the robot helper (to which I was thanked). I also dropped a steak accidentally on the floor, and without hesitation I bent down and picked it up, then threw it across the room. (Pretty much what happens back when I’m in charge of the kitchen anyway.) Turning to my left, I walked towards a fridge which opened automatically with more ingredients inside. I took out a steak, turned to the left again, and using my other controller moved towards a button on the microwave. The door swung open, I placed the steak inside, then I also grabbed a bottle of wine from next to the microwave and put it in as well, before shutting the microwave door how I would normally do so at home. Next to the microwave was a chopping board with a knife and an onion, and when I picked up the knife to cut the onion, it shattered. I couldn’t pick up the pieces of the knife (a safety feature I was told), but I was able to place the onion on a plate and interact with a bell to get the plate taken away.
The second scene put me onto the deck of a sunken ship, surrounded by rotting wood and schools of fish. When I moved around I could see how far down the ship went, but when a fish swam into my face I tried to swat it away, knocking the controller into the headset. After a couple of moments a big whale swam by with the detail on the whale being quite clear. The movement was also smooth, perhaps indicative of 60-90 Hz, but it was not 120 Hz.
The third demo was an example of how the hardware and software can be used in a creative context (I mean more creative than microwaving wine). This was essentially a paint program but allowed the artist to draw in 3D. The right hand was the brush, while using a combination of controls on the left hand allowed the left to act like a palette. I was able to do a full RGB selection similar to most imaging programs, as well as change the type of brush from something more normal to oil paints or a rainbow. I was able to move around in 3D as well, drawing from different perspectives. There was an obligatory drawing of something rude, but then I examined how accurate the software was. It seemed that I could draw with a granularity of about 15mm, especially when drawing long lines. But software like this could be used by developers interested in making and viewing 3D models with immersion and getting a feeling for the size of whatever is being created.
The next demo was best described as a table top game coming to life. I started standing in the middle of the table, watching a faction in a castle fighting off some invaders. In true table-top style everything looked a bit small. But I was able to walk around the scene viewing it from different angles both outside and in the table itself. I was also able to peer inside the 3D models, to get perspectives that would not otherwise be available in the real world.
The final demo was somewhat of a treat. In true Valve fashion I was thrust into a Portal based room with GLaDOS issuing commands over a loud speaker. The artistry was pure Portal through and through, with Portal 2 based figurines. GLaDOS issued commands to open drawers, which the game allowed you to open the wrong drawer twice before opening the right drawer. In the second incorrect drawer was a piece of really old cake, but there were experiments to run and research to be done – thankfully I was still alive so I continued with the demo and opened the right drawer which housed a small Portal stick figure community. Unfortunately GLaDOS told me that looking into the drawer would cause the community to worship me as a god and in the end we had to close the drawer and there was the sound of something burning. I then walked over to the other side of the room and was instructed to open a door via a lever, causing a damaged Portal 2 robot to stumble into the room and be attached to a harness. GLaDOS instructed me to repair the robot by pulling a few parts before giving some incomprehensible instruction that caused the robot to disintegrate. The robot parts were cleared and the walls dismantled to show GLaDOS who expressed disappointment – ‘Oh, it’s you…’. At this point I was on a small segment of floor while I could see far and wide into the Portal underground moving bits around, before the walls around me were quickly rebuilt into what looked like the start of a Portal level. There was even a companion cube.
Thus ended the demo.
The interactivity in each of the demos was great, and it felt natural. Aside from the commands from the HTC guide coming through the headset explaining what was going on (it was a two way conversation, I asked about the hardware), I felt immersed. I could just about see the pixels in the set, but the two controller method for some of the tasks felt like the right thing to do. The PC hardware used for the demo was a single GPU system (I was told a current generation high end card), indicating that no massive tri-SLI system is needed at this point. For a couple of the demos (such as the table top), the scene was a little blurry but I have experienced this before on VR demos – on the large scale everything seems great, but some of the low resolution demos on small models can cause some defocusing. That being said, the Portal demo was filled with detail and I had no issues there, even though there were some intricate detail within the models.
It was difficult to ascertain what kind of technology the display was using due to the lack of consistent black areas, but there was no discernable backlight bleed during gameplay. Unlike Oculus, although I didn’t specifically think about it during the demo there seemed to be a lack of directional audio, as in computation based on hard direction, but I may be mistaken. I felt I was in the moment, and although there was not some massive movement scene similar to the Oculus demo, the package did feel like it could compete.
The headline for this piece ‘Better than Oculus’ is supplanted by a question mark. It is difficult to pick between the two, but if I was forced at gunpoint to spend my own money, I feel it is the HTC model with Steam VR that would get my notes. If not from the hardware perspective, but then the future potential if Valve gets properly onboard with a major title at launch. HTC currently creating units for developers, and is aiming for a 2015 launch. There is talk with regards to several SKUs with the high end one relying on sensors mounted to the wall and perhaps a lower one with gyroscopes. As always with these VR demos, when the devices come to market a key element is going to revolve around content. Here Valve can do serious damage if a big title or two was released along with the headset. The obvious lines are Half-Life 3 and Portal 3, but there was no indication from HTC that anything was going to be mentioned.
The key difference between the HTC and Oculus demos was immersion combined with interactivity. The Oculus demo was technically great, but the interactive elements from HTC along with the Valve input made the experience feel more natural. There are clear hurdles for both implementations, particularly for Oculus on content. Even though Oculus is working with lots of developers, I also inquired about the interoperability between software created for the HTC headset and the Oculus one but the comments pointed towards locking in the software to each device. The HTC/Valve implementation will have to have some bundled content when it goes on general sale and Valve already has the distribution ecosystem in place as well as the gaming franchises to make the world implode.
One of the big announcements to come out of Mobile World Congress this week is regarding HTC’s new VR headset, called the Vive or ‘Re’ Vive, depending on where you see it, and their collaboration with Valve who appears to be providing the software backbone for the hardware. As part of our rush of MWC meetings, HTC invited me (only one person per publication) to experience the Vive developer kit with Steam VR. Having tested the Oculus Rift ‘Crescent Bay’ model at CES and having had a chance to sample Samsung’s Gear VR while here for a few minutes, it is interesting to feel how the VR experiences differ.
At this point in time, HTC is releasing next to no specifications about their developer kit, and the embargo placed today was purely for experiential testing. We were only able to take pictures of the device, and not the room in which it was being tested. Thus the following is purely from my perspective of using the kit and the pre-programmed demos that were provided. Overall each person was allowed 20 minutes to cycle through the demos, similar to that from Oculus at CES.
Functionally I was lead into a near empty room several meters square, put on the headset, and a HTC employee placed a set of headphones over my ears. When the demo started up, the guide held up two controllers, one for each hand. Within the headset, I saw the controllers represented on the screen and reached out to them. The headset and controllers both had a number of wires coming from them, but I was told that the end goal is to make the controllers wireless, and the headset would in the final version have only a single mini-HDMI cable to provide the bandwidth for the display.
The two controllers were designed such that one is for each particular hand with a depressed pad on each (kind of like the initial Steam box controller renders), along with a rear trigger and two grip buttons. Within the first demo, depressing the left trigger caused a balloon to appear from the controller, which I could then let go and punch away with either controller. After a few balloons the main demo began.
The first scene was a basic white field to describe the motions of the set up. I could move my head, walk about 2 meters in each direction before hitting an invisible wall (or an actual wall), as well as duck to get a different perspective from close to the ground. The white field was filled with columns moving up and down to get a sense of perspective.
The next scene was interactive – I was in a kitchen in front of a chopping board with vegetables, with a pot on the hob to my right and a robot helper ahead. With some mild instruction, I used the controllers and the triggers to pick up some of the objects. I put some bread into the pot and threw the rolling pin at the robot helper (to which I was thanked). I also dropped a steak accidentally on the floor, and without hesitation I bent down and picked it up, then threw it across the room. (Pretty much what happens back when I’m in charge of the kitchen anyway.) Turning to my left, I walked towards a fridge which opened automatically with more ingredients inside. I took out a steak, turned to the left again, and using my other controller moved towards a button on the microwave. The door swung open, I placed the steak inside, then I also grabbed a bottle of wine from next to the microwave and put it in as well, before shutting the microwave door how I would normally do so at home. Next to the microwave was a chopping board with a knife and an onion, and when I picked up the knife to cut the onion, it shattered. I couldn’t pick up the pieces of the knife (a safety feature I was told), but I was able to place the onion on a plate and interact with a bell to get the plate taken away.
The second scene put me onto the deck of a sunken ship, surrounded by rotting wood and schools of fish. When I moved around I could see how far down the ship went, but when a fish swam into my face I tried to swat it away, knocking the controller into the headset. After a couple of moments a big whale swam by with the detail on the whale being quite clear. The movement was also smooth, perhaps indicative of 60-90 Hz, but it was not 120 Hz.
The third demo was an example of how the hardware and software can be used in a creative context (I mean more creative than microwaving wine). This was essentially a paint program but allowed the artist to draw in 3D. The right hand was the brush, while using a combination of controls on the left hand allowed the left to act like a palette. I was able to do a full RGB selection similar to most imaging programs, as well as change the type of brush from something more normal to oil paints or a rainbow. I was able to move around in 3D as well, drawing from different perspectives. There was an obligatory drawing of something rude, but then I examined how accurate the software was. It seemed that I could draw with a granularity of about 15mm, especially when drawing long lines. But software like this could be used by developers interested in making and viewing 3D models with immersion and getting a feeling for the size of whatever is being created.
The next demo was best described as a table top game coming to life. I started standing in the middle of the table, watching a faction in a castle fighting off some invaders. In true table-top style everything looked a bit small. But I was able to walk around the scene viewing it from different angles both outside and in the table itself. I was also able to peer inside the 3D models, to get perspectives that would not otherwise be available in the real world.
The final demo was somewhat of a treat. In true Valve fashion I was thrust into a Portal based room with GLaDOS issuing commands over a loud speaker. The artistry was pure Portal through and through, with Portal 2 based figurines. GLaDOS issued commands to open drawers, which the game allowed you to open the wrong drawer twice before opening the right drawer. In the second incorrect drawer was a piece of really old cake, but there were experiments to run and research to be done – thankfully I was still alive so I continued with the demo and opened the right drawer which housed a small Portal stick figure community. Unfortunately GLaDOS told me that looking into the drawer would cause the community to worship me as a god and in the end we had to close the drawer and there was the sound of something burning. I then walked over to the other side of the room and was instructed to open a door via a lever, causing a damaged Portal 2 robot to stumble into the room and be attached to a harness. GLaDOS instructed me to repair the robot by pulling a few parts before giving some incomprehensible instruction that caused the robot to disintegrate. The robot parts were cleared and the walls dismantled to show GLaDOS who expressed disappointment – ‘Oh, it’s you…’. At this point I was on a small segment of floor while I could see far and wide into the Portal underground moving bits around, before the walls around me were quickly rebuilt into what looked like the start of a Portal level. There was even a companion cube.
Thus ended the demo.
The interactivity in each of the demos was great, and it felt natural. Aside from the commands from the HTC guide coming through the headset explaining what was going on (it was a two way conversation, I asked about the hardware), I felt immersed. I could just about see the pixels in the set, but the two controller method for some of the tasks felt like the right thing to do. The PC hardware used for the demo was a single GPU system (I was told a current generation high end card), indicating that no massive tri-SLI system is needed at this point. For a couple of the demos (such as the table top), the scene was a little blurry but I have experienced this before on VR demos – on the large scale everything seems great, but some of the low resolution demos on small models can cause some defocusing. That being said, the Portal demo was filled with detail and I had no issues there, even though there were some intricate detail within the models.
It was difficult to ascertain what kind of technology the display was using due to the lack of consistent black areas, but there was no discernable backlight bleed during gameplay. Unlike Oculus, although I didn’t specifically think about it during the demo there seemed to be a lack of directional audio, as in computation based on hard direction, but I may be mistaken. I felt I was in the moment, and although there was not some massive movement scene similar to the Oculus demo, the package did feel like it could compete.
The headline for this piece ‘Better than Oculus’ is supplanted by a question mark. It is difficult to pick between the two, but if I was forced at gunpoint to spend my own money, I feel it is the HTC model with Steam VR that would get my notes. If not from the hardware perspective, but then the future potential if Valve gets properly onboard with a major title at launch. HTC currently creating units for developers, and is aiming for a 2015 launch. There is talk with regards to several SKUs with the high end one relying on sensors mounted to the wall and perhaps a lower one with gyroscopes. As always with these VR demos, when the devices come to market a key element is going to revolve around content. Here Valve can do serious damage if a big title or two was released along with the headset. The obvious lines are Half-Life 3 and Portal 3, but there was no indication from HTC that anything was going to be mentioned.
The key difference between the HTC and Oculus demos was immersion combined with interactivity. The Oculus demo was technically great, but the interactive elements from HTC along with the Valve input made the experience feel more natural. There are clear hurdles for both implementations, particularly for Oculus on content. Even though Oculus is working with lots of developers, I also inquired about the interoperability between software created for the HTC headset and the Oculus one but the comments pointed towards locking in the software to each device. The HTC/Valve implementation will have to have some bundled content when it goes on general sale and Valve already has the distribution ecosystem in place as well as the gaming franchises to make the world implode.
As a rather unexpected move, SanDisk has announced that it will be stepping into the storage array business with its in-house designed InfiniFlash all-flash array series. The driving force behind SanDisk’s strategy is big data as the current solutions on the market don’t meet all the needs that big data customers have. Most importantly the customers are after higher density and lower price as the data centers can easily be hundreds of petabytes in size, so the acquisition cost as well as the power, cooling and property play a major role.
The InfiniFlash comes in three different flavors. The IF100 is an open platform that is designed for OEMs and hyperscale customers with their own software platforms, although it does include a Software Development Kit (SDK) for monitoring purposes. The IF500 is aimed specifically for object storage using ceph (an open source object storage system), which allows very high capacity clusters to be built efficiently. I admit that I need to dig in a bit deeper to fully understand the way object storage works and its benefits, but it requires less metadata than traditional file and block based storage systems do and the overhead of metadata managing is also lower, which practically enables unlimited scale out opportunities. Finally, the IF700 focuses on performance and is built upon Fusion-io’s ION platform with database (Oracle, SAP, etc.) workloads in mind. In a nut shell, the key difference is that the IF100 is hardware only, whereas the IF500 and IF700 come with InfiniFlash OS (I was told it’s Ubuntu 14.04 based). Hence the IF100 is also cheaper at less than $1 per GB, while the IF500 and IF700 will be priced somewhere between $1 and $2 per GB. For an all-flash array that’s a killer price for raw storage (i.e. before compression and de-duplication) as it’s not uncommon for vendors to still charge over $20 per raw gigabyte.
| SanDisk InfiniFlash Specifications | |
| Form Factor | 3U |
| Raw Capacity | 256TB & 512TB |
| Usable Capacity | 245TB & 480TB |
| Performance (IOPS) | >780K IOPS |
| Performance (Throughput) | 7GB/s |
| Power Consumption | 250W (idle) / 750W (active) |
| Connectivity | 8x SAS 2.0 (6Gbps) |
SanDisk offers two capacity configurations: 256TB and 512TB, of which 245TB and 480TB are usable. The ~7% over-provisioning is used mostly for garbage collection with the goal of increasing sustained performance. In terms of performance, SanDisk claims over 780K IOPS, although the presentation showed up to 1 million IOPS at <1ms latency for 4KB random reads with two nodes. The endurance is 75PB of writes, which results in ~4-year lifespan with 90/10 read/write ratio at 6GB/s. For connectivity the InfiniFlash offers eight SAS 6Gbps ports, which can be upgraded to SAS 12Gbps in the future without any downtime.
EDIT: My initial endurance numbers were actually wrong. The endurance for 8KB random writes is 189PB and 2,270TB for sequential 16KB writes. Assuming 6GB/s throughput that would result in a little over 10 years with 8KB random IO with 90/10 read/write ratio.
Architecturally the InifiniFlash is made out of 64 custom 8TB cards. The cards are SAS 6Gbps based, although as usual SanDisk wouldn’t disclose the underlying controller platform. However, judging by the photo it appears to be a Marvell design (see the hint of a Marvell logo on the chip), which honestly wouldn’t surprise me since SanDisk has relied heavily on Marvell for controllers, including SAS ones (e.g. Optimus Eco uses the 88SS9185 SAS 6Gbps controller). As for the NAND, SanDisk is using its own 1Ynm 64Gbit MLC (i.e. second generation 19nm) and each card consists of 64 eight-die NAND packages (this actually sums up to only 4TB, so I’m waiting to hear back if there’s a typo somewhere in the equation — maybe it’s a 128Gbit die after all). SanDisk will also be using different types of NAND in the future ones new nodes become available in volume.
EDIT: As I speculated, the die capacity is 128Gbit (16GB), so each package consists of eight 128Gbit dies that yields 128GB per package and with 64 packages per card the total raw capacity is 8TB.
The InfiniFlash will be sold through the usual channel (i.e. through OEMs, SIs & VARs), but SanDisk is also selling it directly to some of its key hyperscale customers that insisted on working directly with SanDisk. While it may seem like SanDisk is stepping into its partners area with direct sales, the big hyperscale customers are known to built their own designs, so in most cases they wouldn’t be buying from the OEMs, SIs or VARs anyway. Besides, the InfiniFlash doesn’t include any computing resources, meaning that the partners can differentiate by integrating InfiniFlash to servers for an all-in-one approach, whereas the units SanDisk is selling are just a bunch of raw flash.
While the InfitiFlash as a product is certainly exciting and unique, it’s not as interesting as the implications of this strategical move. In the past the SSD OEMs have kept their fingers away from the storage array market and have simply focused on building the drives that power the arrays. This strategy originates from the hard drive era where the separation was clear: companies X made the hard drives and companies Y integrated them to arrays. Since there wasn’t really a way to customize a hard drive in terms of form factor and features, there was no use of merging X and Y given that the two had completely different expertise.
With flash that’s totally different thanks to the versatility of NAND. Only the sky is the limit when it comes to form factors as NAND can easily scale from a tiny eMMC package to 2.5″ drive and most importantly to any custom form factor that is desired. SanDisk’s biggest advantage in the storage array business is without a doubt its NAND expertise. Integrating NAND properly takes much more knowhow and effort than hard drives because ultimately you have to abandon the traditional form factors and go custom. It’s not a surprise that the InfiniFlash is the densest all-flash array on the planet because as a NAND manufacturer SanDisk has the supply and engineering talent to put 8TB behind a single controller with a very space efficient design. The majority of storage array vendors simply source their drives from the likes of SanDisk, Intel and Samsung, so they are limited by the constraining form factors that are available on the open market.
The big question is how will the market react and will this become a trend among other SSD vendors as well. The storage array market is certainly changing and the companies that used to rule the space are starting to lose market share to smaller, yet innovative companies. As the all-flash array market matures, it seems logical for the SSD vendors to step in and eliminate the middle-man for increased vertical integration. The future remains to be seen, but I wouldn’t be surprised to see others following SanDisk’s lead. The bar has definitely been set high, though.