This feels fluff to me on the part of the author (whose work I don’t want to trivialize) but I don’t think they’ve actually looked deeper than a paper spec sheet on this.
1. Yes it has the same number of cores as a 5070 mobile. It’s also running at a shared peak of 2/3 the bandwidth and a shared peak of 2/3 the TDP. The GPU by itself will likely perform at half the dedicated units performance
2. Apple may not have SVE2 but they do have the AMX (private) and SME. I don’t see why he thinks the SVE2 will give him more performance than the SME.
3. He mentions a single core type but doesn’t mention the total makeup. We already have known for a year how the DGX Spark compares to Apple chips. For CPU it’s roughly equivalent to an M3 Pro and for GPU compute (not rasterization) it’s between an M4 Pro and M4 Max without considering bandwidth.
The real advantage to these is that they run CUDA. That’s it. Otherwise when they launch they’ll be 2-3 generations behind where Apple is and 1 gen behind AMD.
The other super power of the DGX Spark was the NIC for pairing them together. But that’s been removed here too.
> GPU compute (not rasterization) it’s between an M4 Pro and M4 Max without considering bandwidth
You are likely thinking about token generation which is dependent on memory bandwidth where Apple has an edge. Spark's GPU compute is way higher than even M5 Max, around 2x FP32 TFlops... It's literally 6144 CUDA cores like desktop 5070, slowed down by slow memory.
It is absolutely fluff, and the only reason this worthless tweet is on the front page of HN is that this audience has a habit of canonizing certain people, and treating each of their bowel movements as prophetic.
Guy suddenly became aware of a chip that the rest of the industry long knew about, seems completely unaware of the competitors, and posts about how it's a BEAST and will be a GAME CHANGER.
Like the DGX Spark was a game changer? Eh, it has mostly been a massive disappointment. An overpriced nvidia laptop isn't going to change the equation an iota.
The 900 GB/s is from the NVLink-C2C interconnect, if you were wondering about that. They quote "up to 900 GB/s of bidirectional bandwidth between GPU and CPU".
Mind you thats not to/from memory, which indeed only has 273 GB/s.
I know this comes up a lot on HN because its not primarily a graphics community but:
1. Gaussian Splats are very expensive to render. They capture a lot of detail which makes them seem cheaper than an equivalent raster render of that quality, but they wouldn't meet real time AAA game performance requirements
2. Gaussian Splats don't have a concrete surface. Want to cast shadows or do physics? It's doable but very tricky. Want to relight them? Also tricky. What is the exact surface point that you want to affect or sample for any particular operation? Deformations also become very difficult to do well.
3. Gaussian Splats are not sharp. You can get sharper with different kernel types or higher density of points, but your costs go up as well.
4. Gaussian splats are awful for any kind of path tracing. You can do it but you go back to the issues above. So mixing and matching traditional content with splats becomes a performance bottleneck.
I don't think you'll see a AAA game use splats for more than something like cinematics in the near term.
I'm working on the vision component of a drone racing stack. Could I use GS to render my living room as a digital playground to train my vision models in?
I know nothing about the technology but the alternative is creating a 3d model of my living room which is also outside my skill-set.
> Could I use GS to render my living room as a digital playground to train my vision models in?
Yes, its what the autonomous car people are doing.
However you might want to do photogrammetry first (https://github.com/alicevision/meshroom opensource) as that produces a mesh which you can use to detect collisions easier. The downside is that transparent objects render really badly. but it is a lot faster to render
Ehhh if it's just for looking and you don't have anything lidar just go for splats they're way better behaved, mostly because they don't need to understand a concept of "surface" they just understand "splat with spherical harmonics of view-dependant color".
Splats would work and it’s what a lot of automation folks use. But the risk is that your splats aren’t tight to the surface and that can cause false positives.
For training you can do a hybrid geometry plus splats workflow. Have geometry that you can constantly raycast against and have as an input to your vision training or to get accurate depth buffers.
The workflow for splats and photogrammetry are very similar.
There wouldn't be any cloud. Splats are still local, but all the lighting and texture are pre-rendered. The problem is they're not interactive, so they'd be good for a lot of the environment but your main character and other things that need to be interactive would need to use a different approach.
I'm curious how the RTX Spark will perform and what it'll cost.
The big things it has going for it are CUDA and DX support...but performance wise this is pretty weak.
It's a 5070 (mobile) equivalent core set but in a <100W envelope for combined components, whereas the full 5070 mobile is more than that by itself.
It also has a very constrained bandwidth in comparison, and a lot of older ARM cores. NVIDIA and Microsoft better be able to be competitive on price because unless you specifically need CUDA or Windows, this just doesn't feel like a very competitive product to a Mac Studio?
Not traditional Gaussian splats since they have their lighting baked in but there are alternate forms that have the ability to do so but largely in academia or very specific use cases
It means that you can lock the end point to a position and figure out the rest of the pose.
Imagine a character walking with forward kinematics. Every time you move the characters hips , you’d have to rotate the leg joints and make sure the foot doesn’t slide. Remember virtual characters don’t have friction.
IK lets you lock the foot in a spot so you can animate the body above it without having to spend time matching the foot position.
In a real world, this is like if you tried to put your hand on a door handle while jumping up and down. It’s easier to keep position when you hold the handle than if you were to just touch it.
It also means you can move the point and the rest of the joints adjust. Want to make your model walk? Just figure out the foot motion and the rest of the details are calculated based on joint range of motion and limb length.
You must be a hoot at parties. Getting to pick and choose what you and everyone else around you should care about and how they define terms to fit exactly what’s important to just you.
Orthography may be independent of language, but shockingly in society, it’s part of written language.
brother you were trying to "gotcha" me by saying my language is imprecise because i didn't capitalize. i'm pointing out that the two things have literally nothing to do with each other.
It’s not trying to be a gotcha, it’s pointing out the sheer arrogance to criticize someone for your own decision to take a commonly used metaphor literally, and then to go on and act high and mighty because you’ve decided to be the outlier in society who’s decided that you get to pick and choose language norms at your whim.
But you do you. I’m sure it gives some self satisfaction to pick pseudo intellectual arguments. There’s no point continuing this discussion because it’s clearly been in bad faith from the get go.
I like how Apple related conversations, I can’t make a point about how restrictive the OS has become and how most users are probably using restricted devices without apologists telling me “not enough people need to disable that stuff so it doesn’t count”. I’m sure you will brag about the merits of Apple’s locked down security of the App store, that’s not a restriction right? It’s just “security”.
Wherever you pundits want to move the goal post I guess.
Nobody is moving the goalposts. You just poorly choose words.
I don’t think you actually understand the problem space of what you’re describing or you wouldn’t call it DRM.
And as such I don’t think you actually have a grasp of what is affected. But you’ve deles that everyone who doesn’t think it’s the same level of issue as you is somehow beneath your intellect based on your other comments.
1. Yes it has the same number of cores as a 5070 mobile. It’s also running at a shared peak of 2/3 the bandwidth and a shared peak of 2/3 the TDP. The GPU by itself will likely perform at half the dedicated units performance
2. Apple may not have SVE2 but they do have the AMX (private) and SME. I don’t see why he thinks the SVE2 will give him more performance than the SME.
3. He mentions a single core type but doesn’t mention the total makeup. We already have known for a year how the DGX Spark compares to Apple chips. For CPU it’s roughly equivalent to an M3 Pro and for GPU compute (not rasterization) it’s between an M4 Pro and M4 Max without considering bandwidth.
The real advantage to these is that they run CUDA. That’s it. Otherwise when they launch they’ll be 2-3 generations behind where Apple is and 1 gen behind AMD.
The other super power of the DGX Spark was the NIC for pairing them together. But that’s been removed here too.
reply