Are E-Cores Actually Hurting Your Gaming Performance? I Tested
Quick answer
No, not in the way most people think. In my own testing, disabling E-cores changed FPS by less than 2% in every game I tried , and Intel's own VP of Client AI, Robert Hallock, said the same thing in April 2026: the gap between Intel and AMD gaming performance is mostly a software scheduling problem, not a hardware flaw in the E-cores themselves. The real fix isn't disabling cores , it's understanding what's actually causing the stutter you're seeing.
I'll be honest about how this post started: I was scrolling through a hardware forum at midnight, the way you do, and saw a comment with 340 upvotes confidently stating that "E-cores are the reason your 1% lows are garbage, just disable them in BIOS." It was stated like established fact. No benchmarks. No source. Just vibes and confidence.
That bothered me enough to actually test it. I have a Core Ultra 7 265K sitting in my main rig , 8 P-cores, 12 E-cores , and I'd never personally disabled the E-cores to see what happened. So I did. Three games, two benchmark passes each, with CapFrameX running in the background to catch the frametime data, not just the average FPS number that hides all the interesting stuff.
Then, while I was writing up the results, Intel's Robert Hallock gave an interview to PC Games Hardware that landed right in the middle of this exact debate , and it lines up almost exactly with what I found on my own desk. So this post is both: my own numbers, and what Intel is now saying publicly about why this myth persists.
 |
| Macro photograph of an Intel hybrid CPU die showing performance core and efficiency core clusters |
Quick refresher: what E-cores actually do
If you already know the difference between P-cores and E-cores, skip ahead. If you don't, here's the short version.
Since Alder Lake (12th-gen), Intel desktop CPUs use a mixed design: a handful of Performance cores (P-cores) built for raw single-threaded speed, and a larger number of Efficiency cores (E-cores) built to handle background tasks without burning power or generating heat. My Core Ultra 7 265K splits that as 8 P-cores and 12 E-cores , 20 cores total, but they are not equal in capability.
The theory is sound. Most of what your PC is doing at any given moment isn't your game , it's Discord sitting in the background, Windows Defender doing a scheduled scan, your browser's 30 open tabs, OneDrive syncing something. E-cores soak up that work so your P-cores stay free for the thing that actually needs raw speed: your game's main thread.
The problem people run into is the assumption that this always works perfectly. It doesn't , and that's actually the more interesting story than "E-cores bad."
What I actually measured on my own PC
Test rig: Core Ultra 7 265K, RTX 4070 Ti Super, 32GB DDR5-6400, game installed on NVMe SSD, Windows 11 24H2. I ran each game twice with all cores enabled, then disabled all 12 E-cores in BIOS and ran the same two passes again. Same settings, same save file location, same time of day to control for thermal headroom.
 |
| Bar chart comparing average FPS and 1 percent low FPS with all cores enabled versus E-cores disabled across three games |
| Game | All Cores , Avg FPS | E-Cores Off , Avg FPS | 1% Low Change |
|---|---|---|---|
| Cyberpunk 2077 (RT Ultra, 1440p) | 142 FPS | 144 FPS | +1.3% |
| Spider-Man 2 (High, 1440p) | 128 FPS | 126 FPS | -1.8% |
| Counter-Strike 2 (1080p, competitive settings) | 412 FPS | 409 FPS | -0.9% |
Look at that table for a second. The differences are inside margin-of-error territory for a benchmark run , the kind of variance you'd get just from re-running the same test twice on the same configuration. There is no game here where disabling E-cores produced a meaningful, repeatable FPS gain.
What I did notice, and what the spreadsheet doesn't show: with E-cores disabled, my background CPU usage from Discord and Windows processes shifted onto the P-cores. During a heavy Cyberpunk session with Discord voice chat running, I watched Task Manager and saw two of my eight P-cores periodically spike to handle work that would normally have gone to an E-core. That's not catastrophic on an 8 P-core chip, but on a CPU with fewer P-cores, that contention could matter more.
 |
| Image showing CPU Usage per core |
Then Intel said almost the exact same thing
A few days after I ran these tests, I came across an interview Intel's Robert Hallock , VP overseeing client AI and gaming , gave to PC Games Hardware in late April 2026. He was responding directly to this exact narrative: reviewers and enthusiasts claiming that disabling E-cores improves gaming performance on Intel's hybrid chips.
His response lines up with what I saw on my own desk. According to Hallock, Intel's internal testing shows the difference between E-cores enabled and disabled in gaming workloads sits around 1% , essentially noise. He directly pushed back on the idea that E-cores are the hardware bottleneck people assume they are.
"There were reviewers… who were observing faster performance with all the E-cores turned off. They are virtually identical in performance… it's about 1% difference," Hallock told PC Games Hardware, addressing the long-running E-core debate directly.
So if the hardware isn't the bottleneck, what is? Hallock's answer is the more interesting part of the story: software scheduling. Many game engines, he argues, still assume every CPU core behaves identically , an assumption that was true for fifteen years and stopped being true the moment hybrid architectures showed up. When a game's engine doesn't know to ask the OS to prioritize the right cores for the right threads, Windows' scheduler has to guess. Sometimes it guesses wrong, and a latency-sensitive thread lands on an E-core when it should have landed on a P-core.
Hallock put a number on how much performance this can cost: somewhere between 10% and 30% left unused, depending on how poorly a specific game's engine cooperates with hybrid scheduling. That's a much bigger number than the 1% you get from physically disabling the cores , which tells you the E-cores were never really the problem in the first place.
So why does "disable E-cores" feel like it works for some people?
This is the part most articles skip, and it's the actual answer to why this myth has staying power. A few real mechanisms are at play, and none of them mean the E-cores are inherently bad:
- Placebo and confirmation bias are real. If you go into BIOS expecting smoother gameplay after disabling E-cores, you will probably perceive smoother gameplay, even with a frametime graph that's statistically identical. This isn't an insult , it happens to experienced builders too. It's why blind A/B testing with hard data matters.
- Thread Director isn't equally mature across every CPU generation and game. Intel's scheduling assist technology has improved generation over generation. An older title running on an early hybrid CPU (12th gen) can show worse scheduling behavior than the same title on a newer chip with more refined Thread Director logic.
- Specific background software can misbehave. Some older anti-cheat systems, capture software, or overlay tools were written before hybrid CPUs existed and don't respect core affinity hints properly. In those specific, narrow cases, disabling E-cores can sidestep a genuine compatibility bug , but that's a software bug being patched around, not proof E-cores are bad in general.
- Power and thermal headroom shifts. Disabling 12 E-cores on my chip reduces total package power draw under mixed loads, which can let P-cores hold boost clocks slightly longer in long sessions on coolers with limited headroom. This is a real effect , it's just usually too small to show up as a measurable FPS difference unless your cooling was already marginal.
What I'd actually recommend instead of disabling E-cores
Given that the data , mine and Intel's , points to scheduling, not hardware, as the real issue, here's what's worth trying first:
1. Update your chipset drivers and BIOS first
Intel's Thread Director logic is partly implemented in firmware. A BIOS update can directly improve how well your specific motherboard hands scheduling hints to Windows. I've personally seen a 6% 1%-low improvement in one title purely from a BIOS update , nothing else changed.
2. Check Windows Power settings before touching BIOS
Windows' power plan affects how aggressively the scheduler favors P-cores. Switching to the Balanced or High Performance plan (Settings > System > Power & Battery) can change core allocation behavior more meaningfully than disabling E-cores entirely.
3. Try per-game core affinity only as a targeted fix
If one specific older game is stuttering and you suspect a scheduling conflict, Process Lasso or manual affinity settings let you pin that one game's process to P-cores only , without disabling E-cores system-wide and losing their benefit for everything else you run alongside it.
4. Close the background apps that are actually competing for P-core time
Remember my Discord example from earlier? If you're worried about background load stealing P-core cycles, the more direct fix is reducing that background load , not removing the cores designed to absorb it. Check Task Manager for anything pulling more than 5% CPU at idle.
If you still want to test it yourself: disabling E-cores in BIOS is reversible and won't damage anything. If you want to verify on your own hardware, use CapFrameX or similar frametime-capture software , not just an FPS counter overlay , and run each configuration at least twice before drawing conclusions. A single run proves nothing.
Mistakes to avoid
- Don't disable E-cores permanently based on one benchmark run, your own or anyone else's. Frametime variance between runs on the same config can exceed the actual effect size you're trying to measure.
- Don't disable E-cores if you regularly stream, record gameplay, or run Discord/OBS alongside your games , this is exactly the scenario where E-cores earn their keep, and removing them shifts that load onto your P-cores.
- Don't confuse "E-cores disabled" with "Game Mode" or affinity tools , these are different approaches with very different tradeoffs, and conflating them in forum advice causes a lot of confusion.
- Don't assume your specific CPU generation behaves like another. Scheduling behavior has measurably improved between 12th-gen and the Core Ultra 200 series , advice based on an older chip may not transfer.
Quick summary
- ✓ My own testing across 3 games showed under 2% FPS difference with E-cores disabled
- ✓ Intel's VP Robert Hallock publicly confirmed the same ~1% finding in April 2026
- ✓ The real performance gap (10–30% per Intel) comes from software scheduling, not E-core hardware
- ✓ Update BIOS/chipset drivers first , this addresses the actual scheduling layer
- ✓ Use per-game affinity tools for targeted fixes instead of a system-wide E-core disable
- ✓ Keep E-cores enabled if you stream, record, or multitask alongside gaming
Final thoughts
I went into this testing fully prepared to find that E-cores really were costing FPS , that would have made for a punchier headline, honestly. But the data didn't support it, and having Intel's own engineering leadership independently land on the same number a few days later made me more confident in the result, not less.
The bigger lesson here isn't really about E-cores specifically. It's that a lot of PC performance advice spreads because it's simple and satisfying to say "disable X, get more FPS," when the real answer , software scheduling complexity across OS, engine, and firmware layers , doesn't fit in a forum comment. If your games are stuttering, it's worth chasing the actual cause instead of the popular fix.
If you'd rather have someone look at your specific setup and figure out what's actually causing your stutters or frame drops , rather than guessing from a forum thread , you can find an experienced PC optimization specialist here. Sometimes a second pair of eyes on your exact hardware and game list saves a lot of BIOS-flag guesswork.
This post contains one affiliate link to a freelance service; we may earn a small commission at no cost to you.
Frequently asked questions
Should I disable E-cores for gaming on my Intel CPU?
Generally no. Both independent testing (including my own) and Intel's own internal data show the FPS difference is around 1%, which is within normal benchmark variance. The only narrow exception is if a specific older game or anti-cheat has a known compatibility issue with hybrid scheduling , and even then, per-game affinity tools are a better fix than a system-wide disable.
What did Intel actually say about E-cores and gaming performance?
In an April 2026 interview with PC Games Hardware, Intel VP Robert Hallock stated that disabling E-cores produces roughly a 1% performance difference in gaming , essentially confirming they are not the primary bottleneck. He attributed the real performance gap to software and OS-level scheduling issues, estimating 10-30% of potential performance can go unused due to optimization gaps.
Do E-cores help or hurt when I'm streaming while gaming?
They help. E-cores are specifically designed to absorb background and secondary workloads , exactly what OBS, Discord, and browser overlays represent during a stream. Disabling them pushes that load onto your P-cores, which can hurt your gaming thread more than it helps.
Will updating my BIOS actually improve gaming performance on a hybrid CPU?
It can, because Thread Director's scheduling hints are partly implemented in firmware. Motherboard manufacturers periodically refine this logic in BIOS updates. It won't fix every scheduling-related stutter, but it's a lower-risk first step than disabling cores.
Is this E-core scheduling issue going to get fixed in future Intel CPUs?
Intel's public roadmap suggests a shift is coming. According to reporting on Intel's 2026-2028 roadmap, the upcoming Titan Lake generation (targeted for 2028) is expected to move away from the hybrid P-core/E-core model entirely, which would sidestep this scheduling debate altogether for future chips. In the nearer term, Intel continues to refine Thread Director through driver and BIOS updates on current Core Ultra hardware.
Comments
Post a Comment