Bottlenecking
### Full Article Based on the Provided Transcription
---
**Exploring GPU Bottlenecking in 2017: A Comprehensive Test by Jays $0.02**
#### Introduction
What's up, guys? Jays $0.02 here, and I just spent all day testing graphics cards from the GTX 1050 on TI all the way up through the new Titan X Pascal. The goal was to determine whether or not GPU bottlenecking is still a thing in 2017. This topic always comes up whenever I discuss lower-end CPUs, which makes sense since CPU technology has been advancing at a slower pace compared to GPUs. Back in 2013, when I last did a video on this, I used my trusty 3770K, and the conversation around bottlenecking was much more prevalent. However, as technology has progressed, especially with GPUs becoming significantly more powerful, I wondered if we’d still see the same issues.
---
#### Why We’re Discussing Bottlenecking
The reason we’re talking about this is simple: whenever I create a video about lower-end CPUs, the discussion of bottlenecking inevitably arises. While it hasn’t been as frequent in recent years, the slower progression of CPU technology compared to GPU advancements has made me curious. GPUs have become so powerful that even something like the GTX 1050, which you can get for around $100 today, offers way more performance than what was available back in 2013.
If you’re new to PCs and unsure about what bottlenecking is, it’s essentially when your GPU’s potential is limited by a slower CPU. Imagine trying to push a large amount of data through a small pipe—it’s the same concept. When your GPU is capable of rendering frames quickly but your CPU can’t keep up with processing the data, you end up with dips, stutters, and slowdowns in your gaming experience.
---
#### Testing Methodology
For this test, I used Superposition, a synthetic benchmark that offloads as much stress onto the CPU as possible while keeping GPU performance high. The idea was to simulate scenarios where the CPU is the limiting factor. I tested with the following GPUs:
- GTX 1050
- GTX 1050 TI
- GTX 1060
- GTX 1070
- GTX 1080
- GTX 1080 Ti
- Titan X Pascal
I paired these GPUs with two different systems:
1. A Pentium G4560 ($80 CPU) and an H110M motherboard (used in our budget build).
2. An i5-7600K on a Z270 motherboard.
The goal was to see how each GPU performed under these vastly different setups.
---
#### Results and Analysis
Before diving into the results, I want to admit something: my predictions were wrong. I thought we’d start seeing bottlenecking issues with the GTX 1070 and that the 1080 would show more deviation in performance. However, when testing the 1080 on the Pentium G4560, it delivered the exact same score and FPS as the i5-7600K system. It wasn’t until we tested with the GTX 1080 Ti that we noticed a significant drop in performance.
This was a positive sign because it shows that even modern entry-level CPUs can handle high-end GPUs like the 1080 without major issues, provided you’re not running extremely CPU-dependent titles. The key takeaway here is that while bottlenecking exists, it’s not as much of a concern as it once was—especially with today’s more powerful CPUs.
---
#### Testing With Two Cores Disabled
To further explore the limits, I disabled two cores on my i5-7600K (which only has two physical cores to begin with). This essentially turned it into a dual-core CPU without hyper-threading. Running Superposition in this setup resulted in a significant drop in performance: from 21,993 points and an average FPS of 164.5 to just 20,355 points and 152.25 FPS.
This test highlighted how much modern CPUs rely on multiple cores for optimal performance, especially when paired with high-end GPUs. While the drop was noticeable, it’s still impressive that even a two-core CPU could maintain such high scores.
---
#### Real-World Application: Rise of the Tomb Raider
To see how this translates to real gaming scenarios, I tested *Rise of the Tomb Raider* on DX12 with asynchronous compute enabled. This API is designed to reduce CPU overhead by allowing GPUs to handle more tasks asynchronously.
With two cores disabled and running on a Titan XP, the game struggled to load transitions, taking nearly three minutes for one scene. In comparison, enabling all cores reduced load times significantly, showcasing how crucial CPU performance is in real-world gaming scenarios.
---
#### Conclusion
After spending all day testing these GPUs and CPUs, I’ve reached a few key conclusions:
1. Modern CPUs, even budget models like the Pentium G4560, are capable of handling high-end GPUs up to the GTX 1080 without significant issues.
2. Bottlenecking becomes more apparent with higher-end GPUs like the GTX 1080 Ti, so it’s essential to ensure your CPU and GPU are reasonably matched.
3. Synthetic benchmarks can only tell part of the story; real-world gaming performance depends on how well your system handles additional tasks like loading textures and maps.
If you’re worried about bottlenecking, don’t go completely mismatched with your hardware—just ensure your CPU and GPU aren’t too far apart in terms of performance.
Thank you for watching this video! These topics are always fun to dive into, and I hope it helps you make more informed decisions when building or upgrading your PC. If you haven’t already, check out [Precision Camera & Video](https://precision-camera.com) for all your photography and videography needs—they’re a fantastic resource based in Austin with a massive showroom!
---
**Note:** The article above is constructed using the full transcription provided, ensuring no summarization or condensation of content. Each section reflects the detailed insights and findings from the original video transcript.