New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Something similar to this.
https://www.hoponenetworks.com/whmcs/index.php?rp=/store/amd-dedicated-servers
AMD Ryzen 7 5700G @ 3.8 Ghz (8 Cores/16 Threads) CPU
64 GB DDR4 RAM
1TB NVMe SSD Storage
$59.95 USD
Would you like to give such a server a try?
) one. I expect that having two smaller servers will be way more affordable than 1 big one.
Then we can see whether one server (with 32GB RAM) is able to handle 4 files (using Intel Quick Sync). If so, we can just add a second (or even third
A trail is possible. Send me a DM if you would like to make use of this offer.
Check your dm
Eight files at once does not define the workload. Share the resolution, codec and actual FFmpeg command, then benchmark the same sample with 1, 2, 4 and 8 jobs.
Compare total frames per second and peak RAM. A queue with fewer concurrent jobs may finish the batch sooner and avoid running out of memory.
That's a good observation and opinion there, already done that though, at least 3-4 at a time not back. I run h265 codec, which why I need reliable dedicated server. Someone already helping with a test server, I hope it yields a good results. God bless that guy
It always will. The downside is just that it is slower.
A GPU might be able to encode eight separate streams of 10-bit 4k@60 video at once but the bitrate for even half-decent quality will be enormous.
Do you even know what you need? how do you plan to transcode your videos? encode is the wrong term for it. You can't transcode 7-8 movies at the same time, you can but it won't be 10x realtime speed. Intel HD Graphics can do 10x 1080p transcoding, and it is what you need for price / performance. Or cough up more money to hetzner second hand with gtx 1080 gpu and it can do a lot more than that. It was 77 euros/month or something.
Yes, but with Quick Sync, 32GB RAM is almost irrelevant for 4 transcodes. The real limit will be the iGPU media engine, codec/profile, resolution, filters, and how many simultaneous encode/decode jobs it can sustain.
Two smaller Intel boxes could actually be a good option here.
I can recommend them, but I had around 15 minutes of downtime last month (I was literally editing the damn kernel.). Very disappointing Also somehow I can push 1Gbps pretty much constantly on a shared line. Again, very disappointing
.
Yep, CPU encoding still wins for quality per bitrate. Nvidia still can’t do it without turning the output into slop. Feels like all of this is just another way to make us pay more for RAM.
The NVENC engine actually doesn't need much RAM. It's most useful for gamers who want to do real-time streaming and simply can't afford anything that takes CPU cycles away from their game. And since the encoder engine is an ASIC separate from the GPU shaders, it doesn't affect the GPU's graphics abilities either. But that's pretty much all it's useful for: High-speed real-time encoding.
That's unlike NVDEC which is better than software because it's faster and produces identical output, and is useful for everyone.
Technically speaking, transcoding is just decoding followed by encoding. And decoding is very fast so it does make sense that he only cares about the encoding part.