3 Years of Service
@uncbuck I think we got a bingo
The best part of that skit was the dog had the money the entire time
Supporter bonus is liveDonate during this limited campaign to receive the advertised supporter incentive.
Limited-time supporter bonus
View the deal
Anti-Social & Lurker - Hideout spot
- Thread starter Axois
- Start date
- Replies 172K
- Views 6M
3 Years of Service
Yes - this one time, and this one time only, listen to Underlord @derdaich:
You must be registered to see attachments
2 Years of Service
Here you go!
Current Build Analysis:
- 4x32GB RAM
Current Spec: 128GB total
Analysis:
This is already an excellent amount of RAM for multitasking and high-performance workloads, including AI image rendering. For AI, especially if you're using deep learning frameworks, 64GB-128GB of RAM is usually more than enough unless you're handling huge datasets. RAM isn't likely to be the limiting factor in your setup unless you have a very specific workload requiring more.
Recommendation:
I don’t see a need to upgrade this unless you're facing specific RAM limitations in AI tasks. If you have spare RAM slots and want to future-proof, you could add 64GB more, but it’s not necessary for general AI rendering or gaming.
- 2x Samsung 990 M.2 4TB SSD
Current Spec: 8TB total NVMe SSD
Analysis:
These are high-performance SSDs, offering very fast read and write speeds (up to 7,450 MB/s). With 8TB of NVMe storage, you have more than enough for gaming, image rendering, and general use. These drives are excellent for reducing load times, increasing data throughput for AI workloads, and ensuring fast storage.
Recommendation:
This is solid. If you're looking to optimize further, you could potentially look at RAID configurations or additional SSDs for backup and redundancy, but for now, your storage seems adequate.
- EVGA 1300 Watt Modular PSU
Current Spec: 1300W, Modular
Analysis:
This is a high-quality power supply with a lot of headroom. Given your current setup (Ryzen 1800X + 1080 Ti), you're not running into any power bottlenecks. Adding a new GPU like the RTX 4080 or 4080 Super should still leave you with ample power.
Recommendation:
You don't need to upgrade this unless you're planning to run multiple high-end GPUs or extreme overclocking setups. For now, 1300W is plenty.
- 2x Asus 28" 4K Monitors
Current Spec: Dual 28" 4K monitors
Analysis:
4K monitors are a fantastic choice for productivity, gaming, and creative work. Two monitors are great for multitasking, and the 4K resolution will give you plenty of screen real estate for rendering and gaming.
Recommendation:
No need to upgrade your monitors unless you're considering something like ultra-wide screens for better multitasking or a higher refresh rate for gaming (if you care about gaming performance). Otherwise, your current 4K setup should be fine.
- Noctua NH-D15 (Air Cooler)
Current Spec: Noctua NH-D15
Analysis:
The NH-D15 is a top-tier air cooler and one of the best on the market. It's very quiet, performs excellently, and is more than enough to handle overclocking or thermal management for your current CPU, the Ryzen 1800X. The only limitation would be if you’re planning on using a CPU with higher power draw or need additional cooling for extreme overclocking.
Recommendation:
If you plan on upgrading to a high-performance CPU like the Ryzen 7000 series or Intel 13th Gen, you may want to reconsider cooling, but the NH-D15 is still highly recommended for most users.
Your Key Concerns & Recommendations:
1. Should you get an RTX 4080 Super?
The RTX 4080 Super is likely going to be an excellent choice for both AI image rendering and gaming. Here's why:- AI Image Rendering: For AI-based tasks, such as those involving large neural networks or image rendering (especially with large datasets), VRAM becomes a crucial factor. The RTX 4080 comes with 16GB of VRAM, which is a significant upgrade over your 1080 Ti (which has 11GB VRAM). The RTX 4080 Super may come with more VRAM (18GB-24GB), which will drastically improve your rendering performance, especially for models with high resolution or those using models like Stable Diffusion, MidJourney, or even large-scale generative models in AI.
The CUDA cores and Tensor cores on the 4080 (or 4080 Super) will also significantly outperform your current 1080 Ti for AI workloads. This is key in speeding up image generation and running more complex models efficiently. You will see a noticeable speed boost and also have room to grow into more demanding AI tasks. - Gaming: The RTX 4080 Super will also crush any gaming task you throw at it. If you're aiming for 4K gaming at ultra settings with high frame rates, the 4080 Super is well beyond the capabilities of your current 1080 Ti. It's excellent for demanding games like Cyberpunk 2077 and Baldur’s Gate 3 at ultra-realistic settings.
Recommendation: Yes, the RTX 4080 Super is worth it for both AI and gaming. It's a massive upgrade from the 1080 Ti in terms of VRAM, GPU cores, and overall performance. If you can afford it, this would be a great investment, especially for future-proofing.
2. What CPU to Pair for AI Image Rendering?
Your current Ryzen 1800X is a great CPU, but it’s a few generations behind. For AI rendering, especially with tasks that benefit from higher core counts and faster single-core performance, you'll want a CPU that pairs well with a powerful GPU.- Option 1: AMD Ryzen 7000 Series (e.g., 7950X)
The Ryzen 7950X is an excellent option for both AI and gaming. It offers 16 cores and 32 threads, which is fantastic for multi-threaded workloads like AI image rendering. It also provides strong single-core performance, which is beneficial for gaming. - Option 2: Intel Core i9-13900K
The i9-13900K offers 24 cores (8 performance cores + 16 efficiency cores), providing a perfect balance of performance for both multi-threaded AI rendering tasks and single-core performance for gaming. It's arguably the best all-around CPU for both gaming and productivity at the moment.
Recommendation:
I’d recommend going for either the Ryzen 7950X or Intel i9-13900K, both of which will give you a significant performance increase in AI tasks and gaming. Pairing these with the RTX 4080 Super will ensure a well-balanced, high-performance system.
Final Build Suggestion:
- CPU: AMD Ryzen 7950X or Intel Core i9-13900K
Both offer excellent performance for both AI and gaming tasks. - GPU: NVIDIA RTX 4080 Super
Massive VRAM and CUDA/Tensor cores for AI rendering, plus top-tier gaming performance. - Motherboard:
- For AMD: ASUS ROG Crosshair X670E Hero
- For Intel: ASUS ROG Z790 Hero
These motherboards support PCIe Gen 5, provide ample RAM support, and offer solid features for high-end CPUs and GPUs.
- RAM: 128GB (4x32GB) of Corsair Vengeance DDR5-6000 (or DDR4 if sticking with older chipsets). Fast RAM will help with AI workloads.
- Storage: Your 2x Samsung 990 M.2 4TB SSDs are already excellent, but ensure you have backups or larger SSDs if you plan on working with massive datasets.
- Cooling: Noctua NH-D15 should be sufficient, but if you switch to Intel, you may also consider a Corsair iCUE H150i Elite Capellix AIO cooler for the i9-13900K if you push overclocking limits.
- PSU: The EVGA 1300W PSU is more than sufficient.
Conclusion:
- RTX 4080 Super is definitely worth it for your needs in AI image rendering and gaming. It’s a massive step up from the 1080 Ti and will provide a smoother experience.
- Pair it with a high-performance CPU like the Ryzen 7950X or Intel i9-13900K to avoid bottlenecks.
- Your existing RAM and SSDs are excellent, but I’d focus on upgrading the CPU/GPU combo for the most noticeable performance gains.
1 Year of Service
Yes, it feels so weird to be a person that never interacts with anything and is now pumping out comments
You must be registered to see media
3 Years of Service
You must be registered to see attachments
You must be registered to see attachments
3 Years of Service
Holy shit. I was not expecting that level of detail.
Thoughts on AMD GPUs? Stick with NVIDIA?
Cooler should be good if I don't plan to overclock?
Thanks again for the detail!
3 Years of Service
There, there...Yes, it feels so weird to be a person that never interacts with anything and is now pumping out comments
You must be registered to see media
You must be registered to see attachments
3 Years of Service
Not really, there are plenty of people who we see rarely because they don't like to talk much. There is a lot of fear even talking, even some of the most prolific users have talked about it.
Luckily there isn't much required to get to what most people want, anything else is fluffYes, it feels so weird to be a person that never interacts with anything and is now pumping out comments
You must be registered to see media
Honestly with how nVidia has been treating the manufacturers lately, my next gpu will be AMD, and I'll probably partner it with an AMD cpu. EVGA completely left the scene due to how nVidia treated them.
On a scale of 1-10, how disgusted were you saying to listen to @Underlord ?
Underlord
The Chosen Lord of Coins Dark Lord of the Sith
Registered
Lewd
Prestige 1
Prestige 4
Prestige 5
Prestige 6
zoidberg
2 Years of Service
Got the cash might as well stay with NVIDIA the apple of GPUs
But RX 7900 XTX and RX 7900 XT are pretty good.
2 Years of Service
When choosing between NVIDIA and AMD for your GPU, it’s essential to consider your specific use case—both for AI rendering and gaming.
NVIDIA GPUs:
- CUDA & Tensor Cores: NVIDIA dominates in AI and machine learning workloads due to its CUDA cores (for general GPU compute) and Tensor cores (for deep learning tasks). This makes NVIDIA GPUs a superior choice if you’re working with frameworks like TensorFlow, PyTorch, or running AI models. NVIDIA's software ecosystem, including CUDA, cuDNN, and OptiX, is optimized for performance in these areas.
- DLSS (Deep Learning Super Sampling): NVIDIA also has proprietary technologies like DLSS for gaming, which helps with performance while maintaining visual quality. If gaming at 4K is a priority (especially for titles like Cyberpunk 2077), this can be a significant advantage.
- Ray Tracing Performance: NVIDIA’s RTX series has better real-time ray tracing performance than AMD’s offerings. If you want the best visuals and performance for ultra-realistic gaming, NVIDIA is still ahead in this area.
AMD GPUs:
- Price-to-Performance: AMD tends to offer better value for money compared to NVIDIA, especially for gaming. In the high-end GPU market, AMD's RX 7000 series (such as the RX 7900 XTX) competes well with NVIDIA’s 4080, often at a more competitive price point. The price/performance ratio can make AMD a very appealing choice for gaming-focused builds.
- OpenCL Support: AMD GPUs generally offer strong performance in OpenCL (a framework similar to CUDA) for general-purpose computing. However, if your AI workloads are based around CUDA-specific optimizations, NVIDIA would still be the go-to.
- Ray Tracing and FidelityFX Super Resolution (FSR): AMD has improved ray tracing performance with its latest GPUs, but it's still behind NVIDIA in this department. FSR (AMD's version of DLSS) can improve performance, but it doesn’t quite match the quality of NVIDIA’s DLSS 2.0 in many cases.
Summary:
- For AI and deep learning, NVIDIA is the clear winner, as its ecosystem is more mature and optimized for these tasks, thanks to CUDA and Tensor cores.
- For gaming, AMD offers strong competition, particularly in price-to-performance. AMD GPUs are great for high-end gaming, but if you're prioritizing ray tracing or DLSS features, NVIDIA would be better.
Should the Noctua NH-D15 be Enough if You Don't Plan to Overclock?
Yes, the Noctua NH-D15 should be more than adequate even if you don't plan to overclock. Here’s why:- Excellent Airflow and Cooling Performance:The NH-D15 is considered one of the best air coolers on the market. It can handle the cooling requirements of high-performance CPUs (including the Ryzen 7950X or Intel i9-13900K) without any issues under stock settings, even in demanding workloads. The dual tower design and the premium Noctua fans ensure very efficient heat dissipation.
- Efficiency Without Overclocking:Even without overclocking, modern high-performance CPUs can still run hot under heavy load (e.g., rendering, gaming, or multitasking). However, the NH-D15 is built to handle those thermal demands and maintain excellent temperatures. Since you’re not planning to overclock, you don’t need to worry about pushing your CPU to thermal limits. The NH-D15 will provide more than enough cooling power at stock speeds.
- Low Noise:One of the biggest selling points of the NH-D15 is its quiet operation. Even under load, it remains relatively silent, which can be a major benefit if you’re working in a quiet environment or value low noise levels.
- Future-Proofing:While the NH-D15 is overkill for stock cooling needs, it does provide a level of future-proofing, especially if you decide to push your CPU a bit (e.g., moderate overclocking) down the line. It's a solid, long-term investment that ensures your CPU stays cool no matter what workload you’re throwing at it.
Summary:
Even without overclocking, the Noctua NH-D15 will easily handle the cooling needs of modern CPUs, including the Ryzen 7950X or Intel i9-13900K, without any thermal issues. It’s one of the best air coolers available and offers near-silent operation. So, unless you’re aiming for extreme overclocking, there’s no need to upgrade the cooler.Final Thoughts:
- GPU: Stick with NVIDIA, specifically the RTX 4080 Super, for superior AI rendering performance (thanks to CUDA and Tensor cores) and gaming (due to ray tracing and DLSS).
- Cooling: The Noctua NH-D15 will be more than enough if you’re not planning to overclock. It’s an excellent choice for keeping your CPU cool and quiet, even under heavy workloads.
U
Try bumping the ram to 64 and git mechanical HDD drive pacific for os and vital documents
3 Years of Service
You must be registered to see attachments
Underlord
The Chosen Lord of Coins Dark Lord of the Sith
Registered
Lewd
Prestige 1
Prestige 4
Prestige 5
Prestige 6
zoidberg
2 Years of Service
4 Years of Service
That looks familiar. Almost as if you've used that one before.. on me.
You must be registered to see attachments
3 Years of Service
3 Years of Service
Where did you find this footage of me!?
4 Years of Service
You must be registered to see attachments
You must be registered to see attachments
See.. I'm not crazy. Well.. yes....I am. Never mind that though!!
Underlord
The Chosen Lord of Coins Dark Lord of the Sith
Registered
Lewd
Prestige 1
Prestige 4
Prestige 5
Prestige 6
zoidberg
2 Years of Service
It is truly a wonderful day best time to buy steam game licenses to play games haha
3 Years of Service
As you hang out and observe, you'll figure it out.
You must be registered to see attachments
2 Years of Service
thank you.
need it.
You must be registered to see attachments
Haha watched only the sneeze. did not know, there was a follow up. But it stays.
Last edited:
Similar threads
Legal Notice
This site provides links to other sites/services, and does not store any files